clients_churn.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div style=\"background: #cceeaa; padding: 5px; border: 1px solid green; border-radius: 5px;\">\n",
    "<font color='green'> \n",
    "    <b><u>КОММЕНТАРИЙ РЕВЬЮЕРА</u></b>\n",
    "<font color='green'><br>\n",
    "<br />\n",
    "Алексей, доброго дня! рад приветствовать тебя в финале) <br />\n",
    "Вот и он подобрался, поздравляю. Остался последний рывок в этом марафоне)<br>\n",
    "<br />\n",
    "Меня зовут Николай. <br />И на этом этапе Вашего движения к новой профессии от меня будут комментарии по написанному коду. <br />Чтобы меньше \"мусорить эфир\" и чтобы максимально наглядно отделяться от основного кода, есть предложение условиться в некоторых обозначениях:<br /> <br />\n",
    "<u><b>ТАКОЙ ШРИФТ</b></u> - всегда начало комментария <br />\n",
    "<font color='green'>такой шрифт</font> - комментарии о том, что всё ОК <br/>\n",
    "<font color='purple'>такой шрифт</font> - комментарии о том, что всё ОК по результату,<br> однако что есть ещё что-то или в коде или есть способы сделать более короткий или быстрый код <br/>\n",
    "<font color='red'>такой шрифт</font> - комментарии о том, что есть критичный момент, влияющий на бизнес-результат проекта.<br/>\n",
    "</font><br /><br />\n",
    "<font color='green'>Как и всегда в спринтах - ставлю своей целью максимально помочь. Поэтому, все мои комментарии - это только способ передать свой опыт для помощи в твоих достижениях."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Описание проекта"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Оператор связи «Ниединогоразрыва.ком» хочет научиться прогнозировать отток клиентов. Если выяснится, что пользователь планирует уйти, ему будут предложены промокоды и специальные условия. Команда оператора собрала персональные данные о некоторых клиентах, информацию об их тарифах и договорах."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Описание данных"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Данные состоят из файлов, полученных из разных источников:\n",
    "\n",
    "- `contract.csv` — информация о договоре;\n",
    "- `personal.csv` — персональные данные клиента;\n",
    "- `internet.csv` — информация об интернет-услугах;\n",
    "- `phone.csv` — информация об услугах телефонии."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Содержание \n",
    "* [Изучение файлов](#analyse_files)\n",
    "* [Вопросы](#questions)\n",
    "* [План](#plan)\n",
    "* [Первичная подготовка данных](#prepare_data)\n",
    "* [Исследовательский анализ данных](#exploratory_data_analysis)\n",
    "* [Подготовка данных для моделей](#prepare_data_for_models)\n",
    "* [Обучение моделей](#models_train)\n",
    "* [Вывод](#conclusion)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np \n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn.model_selection import train_test_split, GridSearchCV\n",
    "from sklearn.metrics import roc_auc_score, roc_curve, accuracy_score\n",
    "from sklearn.utils import shuffle\n",
    "\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "from catboost import CatBoostClassifier\n",
    "\n",
    "from collections import namedtuple\n",
    "\n",
    "RANDOM_STATE = 12345\n",
    "\n",
    "pd.set_option('display.max_columns', None)\n",
    "\n",
    "numeric_columns = ['MonthlyCharges', 'TotalCharges', 'TotalDays']\n",
    "categorial_columns = [\n",
    "    'Type',\n",
    "    'PaperlessBilling',\n",
    "    'PaymentMethod',\n",
    "    'Gender',\n",
    "    'SeniorCitizen',\n",
    "    'Partner',\n",
    "    'Dependents',\n",
    "    'InternetService', \n",
    "    'OnlineSecurity', \n",
    "    'OnlineBackup', \n",
    "    'DeviceProtection', \n",
    "    'TechSupport', \n",
    "    'StreamingTV', \n",
    "    'StreamingMovies', \n",
    "    'MultipleLines',\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def analyze_df(df):\n",
    "    rows_count, columns_count = df.shape\n",
    "    print('Количество стобцов:', columns_count)\n",
    "    print('Количество строк:', rows_count)\n",
    "    display(df.head(5))\n",
    "    display(df.describe())\n",
    "    df.info()\n",
    "\n",
    "def plot_numerical(df):\n",
    "    cols = 3\n",
    "    rows = 2\n",
    "\n",
    "    plt.figure(figsize=(15, 10))\n",
    "\n",
    "    idx = 0\n",
    "    for column in numeric_columns:\n",
    "        idx += 1 \n",
    "    \n",
    "        plt.subplot(rows, cols, idx)\n",
    "        plt.hist(df[column])\n",
    "        plt.title(column)\n",
    "    \n",
    "    for column in numeric_columns:\n",
    "        idx += 1 \n",
    "    \n",
    "        plt.subplot(rows, cols, idx)\n",
    "        plt.boxplot(df[column])\n",
    "        plt.title(column)\n",
    "        \n",
    "    plt.show()\n",
    "\n",
    "    display(df[numeric_columns].describe())\n",
    "    \n",
    "def upsample(features, target, repeat):     \n",
    "    features_zeros = features[target == 0]     \n",
    "    features_ones = features[target == 1]     \n",
    "    target_zeros = target[target == 0]     \n",
    "    target_ones = target[target == 1] \n",
    "\n",
    "    features_upsampled = pd.concat([features_zeros] + [features_ones] * repeat)     \n",
    "    target_upsampled = pd.concat([target_zeros] + [target_ones] * repeat)     \n",
    "    \n",
    "    features_upsampled, target_upsampled = shuffle(         \n",
    "        features_upsampled, target_upsampled, random_state=RANDOM_STATE)          \n",
    "    \n",
    "    return features_upsampled, target_upsampled\n",
    "\n",
    "def plot_roc_curve(target_test, predictions):\n",
    "    print('ROC-AUC score', roc_auc_score(target_test, predictions))\n",
    "    \n",
    "    fpr, tpr, thresholds = roc_curve(target_test, predictions)\n",
    "\n",
    "    plt.figure(figsize=(11, 5))\n",
    "    \n",
    "    plt.plot(fpr, tpr)\n",
    "    \n",
    "    plt.plot([0, 1], [0, 1], linestyle='--')\n",
    "    \n",
    "    plt.xlim([0.0, 1.0])\n",
    "    plt.ylim([0.0, 1.0])\n",
    "    \n",
    "    plt.xlabel('False Positive Rate')\n",
    "    plt.ylabel('True Positive Rate')\n",
    "    \n",
    "    plt.title('ROC-кривая')\n",
    "    \n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div style=\"background: #cceeaa; padding: 5px; border: 1px solid green; border-radius: 5px;\">\n",
    "    <font color='green'> <b><u>КОММЕНТАРИЙ РЕВЬЮЕРА</u></b>\n",
    "</font>\n",
    "<font color='green'><br>с подготовкой - понятно."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"analyse_files\"></a>\n",
    "## Изучение файлов"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "contract_df = pd.read_csv('/datasets/final_provider/contract.csv')\n",
    "personal_df = pd.read_csv('/datasets/final_provider/personal.csv')\n",
    "internet_df = pd.read_csv('/datasets/final_provider/internet.csv')\n",
    "phone_df = pd.read_csv('/datasets/final_provider/phone.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Количество стобцов: 8\n",
      "Количество строк: 7043\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>customerID</th>\n",
       "      <th>BeginDate</th>\n",
       "      <th>EndDate</th>\n",
       "      <th>Type</th>\n",
       "      <th>PaperlessBilling</th>\n",
       "      <th>PaymentMethod</th>\n",
       "      <th>MonthlyCharges</th>\n",
       "      <th>TotalCharges</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>7590-VHVEG</td>\n",
       "      <td>2020-01-01</td>\n",
       "      <td>No</td>\n",
       "      <td>Month-to-month</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Electronic check</td>\n",
       "      <td>29.85</td>\n",
       "      <td>29.85</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>5575-GNVDE</td>\n",
       "      <td>2017-04-01</td>\n",
       "      <td>No</td>\n",
       "      <td>One year</td>\n",
       "      <td>No</td>\n",
       "      <td>Mailed check</td>\n",
       "      <td>56.95</td>\n",
       "      <td>1889.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>3668-QPYBK</td>\n",
       "      <td>2019-10-01</td>\n",
       "      <td>2019-12-01 00:00:00</td>\n",
       "      <td>Month-to-month</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Mailed check</td>\n",
       "      <td>53.85</td>\n",
       "      <td>108.15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>7795-CFOCW</td>\n",
       "      <td>2016-05-01</td>\n",
       "      <td>No</td>\n",
       "      <td>One year</td>\n",
       "      <td>No</td>\n",
       "      <td>Bank transfer (automatic)</td>\n",
       "      <td>42.30</td>\n",
       "      <td>1840.75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>9237-HQITU</td>\n",
       "      <td>2019-09-01</td>\n",
       "      <td>2019-11-01 00:00:00</td>\n",
       "      <td>Month-to-month</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Electronic check</td>\n",
       "      <td>70.70</td>\n",
       "      <td>151.65</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   customerID   BeginDate              EndDate            Type  \\\n",
       "0  7590-VHVEG  2020-01-01                   No  Month-to-month   \n",
       "1  5575-GNVDE  2017-04-01                   No        One year   \n",
       "2  3668-QPYBK  2019-10-01  2019-12-01 00:00:00  Month-to-month   \n",
       "3  7795-CFOCW  2016-05-01                   No        One year   \n",
       "4  9237-HQITU  2019-09-01  2019-11-01 00:00:00  Month-to-month   \n",
       "\n",
       "  PaperlessBilling              PaymentMethod  MonthlyCharges TotalCharges  \n",
       "0              Yes           Electronic check           29.85        29.85  \n",
       "1               No               Mailed check           56.95       1889.5  \n",
       "2              Yes               Mailed check           53.85       108.15  \n",
       "3               No  Bank transfer (automatic)           42.30      1840.75  \n",
       "4              Yes           Electronic check           70.70       151.65  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MonthlyCharges</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>count</td>\n",
       "      <td>7043.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>mean</td>\n",
       "      <td>64.761692</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>std</td>\n",
       "      <td>30.090047</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>min</td>\n",
       "      <td>18.250000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>25%</td>\n",
       "      <td>35.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>50%</td>\n",
       "      <td>70.350000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>75%</td>\n",
       "      <td>89.850000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>max</td>\n",
       "      <td>118.750000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       MonthlyCharges\n",
       "count     7043.000000\n",
       "mean        64.761692\n",
       "std         30.090047\n",
       "min         18.250000\n",
       "25%         35.500000\n",
       "50%         70.350000\n",
       "75%         89.850000\n",
       "max        118.750000"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 7043 entries, 0 to 7042\n",
      "Data columns (total 8 columns):\n",
      "customerID          7043 non-null object\n",
      "BeginDate           7043 non-null object\n",
      "EndDate             7043 non-null object\n",
      "Type                7043 non-null object\n",
      "PaperlessBilling    7043 non-null object\n",
      "PaymentMethod       7043 non-null object\n",
      "MonthlyCharges      7043 non-null float64\n",
      "TotalCharges        7043 non-null object\n",
      "dtypes: float64(1), object(7)\n",
      "memory usage: 440.3+ KB\n"
     ]
    }
   ],
   "source": [
    "analyze_df(contract_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Electronic check             2365\n",
       "Mailed check                 1612\n",
       "Bank transfer (automatic)    1544\n",
       "Credit card (automatic)      1522\n",
       "Name: PaymentMethod, dtype: int64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "contract_df['PaymentMethod'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Month-to-month    3875\n",
       "Two year          1695\n",
       "One year          1473\n",
       "Name: Type, dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "contract_df['Type'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Выводы**:\n",
    "* В данных нет пропусков\n",
    "* Надо переименовать `customerID` -> `CustomerID`\n",
    "* `BeginDate` и `EndDate` надо перевести в формат даты \n",
    "* `No` в `EndDate` означает, что пользователь продолжает пользоваться услугами - это целевой признак \n",
    "* После выделения целевого признака надо будет заменить `No` на `01.02.2020` в `EndDate`\n",
    "* `TotalCharges` надо перевести в числовой тип\n",
    "* К `PaymentMethod`, `Type` и `PaperlessBilling` надо применить One Hot Encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Количество стобцов: 5\n",
      "Количество строк: 7043\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>customerID</th>\n",
       "      <th>gender</th>\n",
       "      <th>SeniorCitizen</th>\n",
       "      <th>Partner</th>\n",
       "      <th>Dependents</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>7590-VHVEG</td>\n",
       "      <td>Female</td>\n",
       "      <td>0</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>5575-GNVDE</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>3668-QPYBK</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>7795-CFOCW</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>9237-HQITU</td>\n",
       "      <td>Female</td>\n",
       "      <td>0</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   customerID  gender  SeniorCitizen Partner Dependents\n",
       "0  7590-VHVEG  Female              0     Yes         No\n",
       "1  5575-GNVDE    Male              0      No         No\n",
       "2  3668-QPYBK    Male              0      No         No\n",
       "3  7795-CFOCW    Male              0      No         No\n",
       "4  9237-HQITU  Female              0      No         No"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SeniorCitizen</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>count</td>\n",
       "      <td>7043.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>mean</td>\n",
       "      <td>0.162147</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>std</td>\n",
       "      <td>0.368612</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>min</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>25%</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>50%</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>75%</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>max</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       SeniorCitizen\n",
       "count    7043.000000\n",
       "mean        0.162147\n",
       "std         0.368612\n",
       "min         0.000000\n",
       "25%         0.000000\n",
       "50%         0.000000\n",
       "75%         0.000000\n",
       "max         1.000000"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 7043 entries, 0 to 7042\n",
      "Data columns (total 5 columns):\n",
      "customerID       7043 non-null object\n",
      "gender           7043 non-null object\n",
      "SeniorCitizen    7043 non-null int64\n",
      "Partner          7043 non-null object\n",
      "Dependents       7043 non-null object\n",
      "dtypes: int64(1), object(4)\n",
      "memory usage: 275.2+ KB\n"
     ]
    }
   ],
   "source": [
    "analyze_df(personal_df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Выводы**:\n",
    "* В данных нет пропусков\n",
    "* Надо переименовать `customerID` -> `CustomerID`, `gender` -> `Gender`\n",
    "* К `gender`, `Partner` и `Dependents` надо применить One Hot Encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Количество стобцов: 8\n",
      "Количество строк: 5517\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>customerID</th>\n",
       "      <th>InternetService</th>\n",
       "      <th>OnlineSecurity</th>\n",
       "      <th>OnlineBackup</th>\n",
       "      <th>DeviceProtection</th>\n",
       "      <th>TechSupport</th>\n",
       "      <th>StreamingTV</th>\n",
       "      <th>StreamingMovies</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>7590-VHVEG</td>\n",
       "      <td>DSL</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>5575-GNVDE</td>\n",
       "      <td>DSL</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>3668-QPYBK</td>\n",
       "      <td>DSL</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>7795-CFOCW</td>\n",
       "      <td>DSL</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>9237-HQITU</td>\n",
       "      <td>Fiber optic</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   customerID InternetService OnlineSecurity OnlineBackup DeviceProtection  \\\n",
       "0  7590-VHVEG             DSL             No          Yes               No   \n",
       "1  5575-GNVDE             DSL            Yes           No              Yes   \n",
       "2  3668-QPYBK             DSL            Yes          Yes               No   \n",
       "3  7795-CFOCW             DSL            Yes           No              Yes   \n",
       "4  9237-HQITU     Fiber optic             No           No               No   \n",
       "\n",
       "  TechSupport StreamingTV StreamingMovies  \n",
       "0          No          No              No  \n",
       "1          No          No              No  \n",
       "2          No          No              No  \n",
       "3         Yes          No              No  \n",
       "4          No          No              No  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>customerID</th>\n",
       "      <th>InternetService</th>\n",
       "      <th>OnlineSecurity</th>\n",
       "      <th>OnlineBackup</th>\n",
       "      <th>DeviceProtection</th>\n",
       "      <th>TechSupport</th>\n",
       "      <th>StreamingTV</th>\n",
       "      <th>StreamingMovies</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>count</td>\n",
       "      <td>5517</td>\n",
       "      <td>5517</td>\n",
       "      <td>5517</td>\n",
       "      <td>5517</td>\n",
       "      <td>5517</td>\n",
       "      <td>5517</td>\n",
       "      <td>5517</td>\n",
       "      <td>5517</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>unique</td>\n",
       "      <td>5517</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>top</td>\n",
       "      <td>2135-DQWAQ</td>\n",
       "      <td>Fiber optic</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>freq</td>\n",
       "      <td>1</td>\n",
       "      <td>3096</td>\n",
       "      <td>3498</td>\n",
       "      <td>3088</td>\n",
       "      <td>3095</td>\n",
       "      <td>3473</td>\n",
       "      <td>2810</td>\n",
       "      <td>2785</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        customerID InternetService OnlineSecurity OnlineBackup  \\\n",
       "count         5517            5517           5517         5517   \n",
       "unique        5517               2              2            2   \n",
       "top     2135-DQWAQ     Fiber optic             No           No   \n",
       "freq             1            3096           3498         3088   \n",
       "\n",
       "       DeviceProtection TechSupport StreamingTV StreamingMovies  \n",
       "count              5517        5517        5517            5517  \n",
       "unique                2           2           2               2  \n",
       "top                  No          No          No              No  \n",
       "freq               3095        3473        2810            2785  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 5517 entries, 0 to 5516\n",
      "Data columns (total 8 columns):\n",
      "customerID          5517 non-null object\n",
      "InternetService     5517 non-null object\n",
      "OnlineSecurity      5517 non-null object\n",
      "OnlineBackup        5517 non-null object\n",
      "DeviceProtection    5517 non-null object\n",
      "TechSupport         5517 non-null object\n",
      "StreamingTV         5517 non-null object\n",
      "StreamingMovies     5517 non-null object\n",
      "dtypes: object(8)\n",
      "memory usage: 344.9+ KB\n"
     ]
    }
   ],
   "source": [
    "analyze_df(internet_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Fiber optic    3096\n",
       "DSL            2421\n",
       "Name: InternetService, dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "internet_df['InternetService'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Выводы**:\n",
    "* В данных нет пропусков\n",
    "* Данные есть не по всем клиентам\n",
    "* Надо переименовать `customerID` -> `CustomerID`\n",
    "* К `InternetService`, `OnlineSecurity`, `OnlineBackup`, `DeviceProtection`, `TechSupport`, `StreamingTV`, `StreamingMovies` надо применить One Hot Encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Количество стобцов: 2\n",
      "Количество строк: 6361\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>customerID</th>\n",
       "      <th>MultipleLines</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>5575-GNVDE</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>3668-QPYBK</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>9237-HQITU</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>9305-CDSKC</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>1452-KIOVK</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   customerID MultipleLines\n",
       "0  5575-GNVDE            No\n",
       "1  3668-QPYBK            No\n",
       "2  9237-HQITU            No\n",
       "3  9305-CDSKC           Yes\n",
       "4  1452-KIOVK           Yes"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>customerID</th>\n",
       "      <th>MultipleLines</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>count</td>\n",
       "      <td>6361</td>\n",
       "      <td>6361</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>unique</td>\n",
       "      <td>6361</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>top</td>\n",
       "      <td>1763-WQFUK</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>freq</td>\n",
       "      <td>1</td>\n",
       "      <td>3390</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        customerID MultipleLines\n",
       "count         6361          6361\n",
       "unique        6361             2\n",
       "top     1763-WQFUK            No\n",
       "freq             1          3390"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 6361 entries, 0 to 6360\n",
      "Data columns (total 2 columns):\n",
      "customerID       6361 non-null object\n",
      "MultipleLines    6361 non-null object\n",
      "dtypes: object(2)\n",
      "memory usage: 99.5+ KB\n"
     ]
    }
   ],
   "source": [
    "analyze_df(phone_df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Выводы**:\n",
    "* В данных нет пропусков\n",
    "* Данные есть не по всем клиентам\n",
    "* Надо переименовать `customerID` -> `CustomerID`\n",
    "* К `MultipleLines` надо применить One Hot Encoding"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"questions\"></a>\n",
    "## Вопросы"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Целевой признак - если `EndDate` равен `No`? \n",
    "* Какую метрику надо использовать? \n",
    "* На какие значения метрики надо ориентироваться? \n",
    "* Можно ли ко всем категориальным значениям применить One Hot Encoding, поскольку категорий мало? "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div style=\"background: #cceeaa; padding: 5px; border: 1px solid green; border-radius: 5px;\">\n",
    "    <font color='green'> <b><u>КОММЕНТАРИЙ РЕВЬЮЕРА</u></b>\n",
    "</font>\n",
    "<font color='green'><br>ОК, данные на базе) Судя по комментариям - тебе удалось глянуть их внимательно!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"plan\"></a>\n",
    "## План"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Изучить данные в файлах - проверить типы, есть ли пропуски, надо ли преобразовывать данные \n",
    "* Подготовить данных на основе выводов, полученных на предыдущем шаге \n",
    "* Провести исследовательский анализ данных - вузализация данных, избавление от выбросов \n",
    "* Перебор моделей и подбор их параметров \n",
    "* Вывод по итогам работы"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"prepare_data\"></a>\n",
    "## Первичная подготовка данных"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Переименуем столбцы, для того, чтобы они все были в одном формате "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['CustomerID' 'BeginDate' 'EndDate' 'Type' 'PaperlessBilling'\n",
      " 'PaymentMethod' 'MonthlyCharges' 'TotalCharges']\n"
     ]
    }
   ],
   "source": [
    "contract_df = contract_df.rename(columns={'customerID': 'CustomerID'})\n",
    "print(contract_df.columns.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['CustomerID' 'Gender' 'SeniorCitizen' 'Partner' 'Dependents']\n"
     ]
    }
   ],
   "source": [
    "personal_df = personal_df.rename(columns={'customerID': 'CustomerID', 'gender': 'Gender'})\n",
    "print(personal_df.columns.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['CustomerID' 'InternetService' 'OnlineSecurity' 'OnlineBackup'\n",
      " 'DeviceProtection' 'TechSupport' 'StreamingTV' 'StreamingMovies']\n"
     ]
    }
   ],
   "source": [
    "internet_df = internet_df.rename(columns={'customerID': 'CustomerID'})\n",
    "print(internet_df.columns.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['CustomerID' 'MultipleLines']\n"
     ]
    }
   ],
   "source": [
    "phone_df = phone_df.rename(columns={'customerID': 'CustomerID'})\n",
    "print(phone_df.columns.values)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Добавим целевой признак `Churn`, который будет означать отказался клиент от услуг или нет.   \n",
    "Если значение в столбце `EndDate` равно \"No\", то значит клиент не ушел и целевой признак равен 0.  \n",
    "Если дата ухода есть, то целевой признак равен 1."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>CustomerID</th>\n",
       "      <th>BeginDate</th>\n",
       "      <th>EndDate</th>\n",
       "      <th>Type</th>\n",
       "      <th>PaperlessBilling</th>\n",
       "      <th>PaymentMethod</th>\n",
       "      <th>MonthlyCharges</th>\n",
       "      <th>TotalCharges</th>\n",
       "      <th>Churn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>7590-VHVEG</td>\n",
       "      <td>2020-01-01</td>\n",
       "      <td>No</td>\n",
       "      <td>Month-to-month</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Electronic check</td>\n",
       "      <td>29.85</td>\n",
       "      <td>29.85</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>5575-GNVDE</td>\n",
       "      <td>2017-04-01</td>\n",
       "      <td>No</td>\n",
       "      <td>One year</td>\n",
       "      <td>No</td>\n",
       "      <td>Mailed check</td>\n",
       "      <td>56.95</td>\n",
       "      <td>1889.5</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>3668-QPYBK</td>\n",
       "      <td>2019-10-01</td>\n",
       "      <td>2019-12-01 00:00:00</td>\n",
       "      <td>Month-to-month</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Mailed check</td>\n",
       "      <td>53.85</td>\n",
       "      <td>108.15</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>7795-CFOCW</td>\n",
       "      <td>2016-05-01</td>\n",
       "      <td>No</td>\n",
       "      <td>One year</td>\n",
       "      <td>No</td>\n",
       "      <td>Bank transfer (automatic)</td>\n",
       "      <td>42.30</td>\n",
       "      <td>1840.75</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>9237-HQITU</td>\n",
       "      <td>2019-09-01</td>\n",
       "      <td>2019-11-01 00:00:00</td>\n",
       "      <td>Month-to-month</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Electronic check</td>\n",
       "      <td>70.70</td>\n",
       "      <td>151.65</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   CustomerID   BeginDate              EndDate            Type  \\\n",
       "0  7590-VHVEG  2020-01-01                   No  Month-to-month   \n",
       "1  5575-GNVDE  2017-04-01                   No        One year   \n",
       "2  3668-QPYBK  2019-10-01  2019-12-01 00:00:00  Month-to-month   \n",
       "3  7795-CFOCW  2016-05-01                   No        One year   \n",
       "4  9237-HQITU  2019-09-01  2019-11-01 00:00:00  Month-to-month   \n",
       "\n",
       "  PaperlessBilling              PaymentMethod  MonthlyCharges TotalCharges  \\\n",
       "0              Yes           Electronic check           29.85        29.85   \n",
       "1               No               Mailed check           56.95       1889.5   \n",
       "2              Yes               Mailed check           53.85       108.15   \n",
       "3               No  Bank transfer (automatic)           42.30      1840.75   \n",
       "4              Yes           Electronic check           70.70       151.65   \n",
       "\n",
       "   Churn  \n",
       "0      0  \n",
       "1      0  \n",
       "2      1  \n",
       "3      0  \n",
       "4      1  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "contract_df['Churn'] = (contract_df['EndDate'] != 'No').astype(int)\n",
    "contract_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div style=\"background: #cceeaa; padding: 5px; border: 1px solid green; border-radius: 5px;\">\n",
    "    <font color='green'> <b><u>КОММЕНТАРИЙ РЕВЬЮЕРА</u></b>\n",
    "</font>\n",
    "<font color='green'><br>Ок, с целевой - понятно."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Заменим все значения \"No\" в `EndDate` на дату выгрузки датасета (01.02.2020) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n"
     ]
    }
   ],
   "source": [
    "contract_df['EndDate'] = contract_df['EndDate'].replace(['No'], ['2020-02-01'])\n",
    "print(contract_df[contract_df['EndDate'] == 'No'].shape[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Переведем `EndDate` и `BeginDate` в формат даты "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 7043 entries, 0 to 7042\n",
      "Data columns (total 9 columns):\n",
      "CustomerID          7043 non-null object\n",
      "BeginDate           7043 non-null datetime64[ns]\n",
      "EndDate             7043 non-null datetime64[ns]\n",
      "Type                7043 non-null object\n",
      "PaperlessBilling    7043 non-null object\n",
      "PaymentMethod       7043 non-null object\n",
      "MonthlyCharges      7043 non-null float64\n",
      "TotalCharges        7043 non-null object\n",
      "Churn               7043 non-null int64\n",
      "dtypes: datetime64[ns](2), float64(1), int64(1), object(5)\n",
      "memory usage: 495.3+ KB\n"
     ]
    }
   ],
   "source": [
    "contract_df['BeginDate'] = pd.to_datetime(contract_df['BeginDate'], format='%Y-%m-%d')\n",
    "contract_df['EndDate'] = pd.to_datetime(contract_df['EndDate'], format='%Y-%m-%d')\n",
    "contract_df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Заведем новый столбец `TotalDays`, который будет означать, сколько пользователь пользовался услугами "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "contract_df['TotalDays'] = (contract_df['EndDate'] - contract_df['BeginDate']).dt.days"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div style=\"background: #cceeaa; padding: 5px; border: 1px solid green; border-radius: 5px;\">\n",
    "    <font color='green'> <b><u>КОММЕНТАРИЙ РЕВЬЮЕРА</u></b>\n",
    "</font>\n",
    "<font color='green'><br>интересная идея!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Переведем `TotalCharges` в числовой тип "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 7043 entries, 0 to 7042\n",
      "Data columns (total 10 columns):\n",
      "CustomerID          7043 non-null object\n",
      "BeginDate           7043 non-null datetime64[ns]\n",
      "EndDate             7043 non-null datetime64[ns]\n",
      "Type                7043 non-null object\n",
      "PaperlessBilling    7043 non-null object\n",
      "PaymentMethod       7043 non-null object\n",
      "MonthlyCharges      7043 non-null float64\n",
      "TotalCharges        7032 non-null float64\n",
      "Churn               7043 non-null int64\n",
      "TotalDays           7043 non-null int64\n",
      "dtypes: datetime64[ns](2), float64(2), int64(2), object(4)\n",
      "memory usage: 550.4+ KB\n"
     ]
    }
   ],
   "source": [
    "contract_df['TotalCharges'] = pd.to_numeric(contract_df['TotalCharges'], errors ='coerce')\n",
    "contract_df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Видно, что в столбце `TotalCharges` появились пропуски, посмотрим на строки с ними "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>CustomerID</th>\n",
       "      <th>BeginDate</th>\n",
       "      <th>EndDate</th>\n",
       "      <th>Type</th>\n",
       "      <th>PaperlessBilling</th>\n",
       "      <th>PaymentMethod</th>\n",
       "      <th>MonthlyCharges</th>\n",
       "      <th>TotalCharges</th>\n",
       "      <th>Churn</th>\n",
       "      <th>TotalDays</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>488</td>\n",
       "      <td>4472-LVYGI</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>Two year</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Bank transfer (automatic)</td>\n",
       "      <td>52.55</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>753</td>\n",
       "      <td>3115-CZMZD</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>Two year</td>\n",
       "      <td>No</td>\n",
       "      <td>Mailed check</td>\n",
       "      <td>20.25</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>936</td>\n",
       "      <td>5709-LVOEQ</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>Two year</td>\n",
       "      <td>No</td>\n",
       "      <td>Mailed check</td>\n",
       "      <td>80.85</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1082</td>\n",
       "      <td>4367-NUYAO</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>Two year</td>\n",
       "      <td>No</td>\n",
       "      <td>Mailed check</td>\n",
       "      <td>25.75</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1340</td>\n",
       "      <td>1371-DWPAZ</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>Two year</td>\n",
       "      <td>No</td>\n",
       "      <td>Credit card (automatic)</td>\n",
       "      <td>56.05</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3331</td>\n",
       "      <td>7644-OMVMY</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>Two year</td>\n",
       "      <td>No</td>\n",
       "      <td>Mailed check</td>\n",
       "      <td>19.85</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3826</td>\n",
       "      <td>3213-VVOLG</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>Two year</td>\n",
       "      <td>No</td>\n",
       "      <td>Mailed check</td>\n",
       "      <td>25.35</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4380</td>\n",
       "      <td>2520-SGTTA</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>Two year</td>\n",
       "      <td>No</td>\n",
       "      <td>Mailed check</td>\n",
       "      <td>20.00</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5218</td>\n",
       "      <td>2923-ARZLG</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>One year</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Mailed check</td>\n",
       "      <td>19.70</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6670</td>\n",
       "      <td>4075-WKNIU</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>Two year</td>\n",
       "      <td>No</td>\n",
       "      <td>Mailed check</td>\n",
       "      <td>73.35</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6754</td>\n",
       "      <td>2775-SEFEE</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>Two year</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Bank transfer (automatic)</td>\n",
       "      <td>61.90</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      CustomerID  BeginDate    EndDate      Type PaperlessBilling  \\\n",
       "488   4472-LVYGI 2020-02-01 2020-02-01  Two year              Yes   \n",
       "753   3115-CZMZD 2020-02-01 2020-02-01  Two year               No   \n",
       "936   5709-LVOEQ 2020-02-01 2020-02-01  Two year               No   \n",
       "1082  4367-NUYAO 2020-02-01 2020-02-01  Two year               No   \n",
       "1340  1371-DWPAZ 2020-02-01 2020-02-01  Two year               No   \n",
       "3331  7644-OMVMY 2020-02-01 2020-02-01  Two year               No   \n",
       "3826  3213-VVOLG 2020-02-01 2020-02-01  Two year               No   \n",
       "4380  2520-SGTTA 2020-02-01 2020-02-01  Two year               No   \n",
       "5218  2923-ARZLG 2020-02-01 2020-02-01  One year              Yes   \n",
       "6670  4075-WKNIU 2020-02-01 2020-02-01  Two year               No   \n",
       "6754  2775-SEFEE 2020-02-01 2020-02-01  Two year              Yes   \n",
       "\n",
       "                  PaymentMethod  MonthlyCharges  TotalCharges  Churn  \\\n",
       "488   Bank transfer (automatic)           52.55           NaN      0   \n",
       "753                Mailed check           20.25           NaN      0   \n",
       "936                Mailed check           80.85           NaN      0   \n",
       "1082               Mailed check           25.75           NaN      0   \n",
       "1340    Credit card (automatic)           56.05           NaN      0   \n",
       "3331               Mailed check           19.85           NaN      0   \n",
       "3826               Mailed check           25.35           NaN      0   \n",
       "4380               Mailed check           20.00           NaN      0   \n",
       "5218               Mailed check           19.70           NaN      0   \n",
       "6670               Mailed check           73.35           NaN      0   \n",
       "6754  Bank transfer (automatic)           61.90           NaN      0   \n",
       "\n",
       "      TotalDays  \n",
       "488           0  \n",
       "753           0  \n",
       "936           0  \n",
       "1082          0  \n",
       "1340          0  \n",
       "3331          0  \n",
       "3826          0  \n",
       "4380          0  \n",
       "5218          0  \n",
       "6670          0  \n",
       "6754          0  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "contract_df[contract_df['TotalCharges'].isnull()]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Видно, что все эти клиенты подключены всего 1 день, заполним пропуски в `TotalCharges` на 0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div style=\"background: #cceeaa; padding: 5px; border: 1px solid green; border-radius: 5px;\">\n",
    "    <font color='green'> <b><u>КОММЕНТАРИЙ РЕВЬЮЕРА</u></b>\n",
    "</font>\n",
    "<font color='green'><br>да, верно, это новички. Хорошо раскрутил)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "contract_df['TotalCharges'] = contract_df['TotalCharges'].fillna(0)\n",
    "contract_df['TotalCharges'].isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Объединим все данные в один датафрейм "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "churn_df = contract_df \\\n",
    "    .merge(personal_df, how='left', on='CustomerID') \\\n",
    "    .merge(internet_df, how='left', on='CustomerID') \\\n",
    "    .merge(phone_df, how='left', on='CustomerID') "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Проверим, что все значения `CustomerID` уникальные "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "churn_df['CustomerID'].is_unique"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Установим `CustomerID` в качестве индекса"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "churn_df = churn_df.set_index('CustomerID')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Количество стобцов: 21\n",
      "Количество строк: 7043\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>BeginDate</th>\n",
       "      <th>EndDate</th>\n",
       "      <th>Type</th>\n",
       "      <th>PaperlessBilling</th>\n",
       "      <th>PaymentMethod</th>\n",
       "      <th>MonthlyCharges</th>\n",
       "      <th>TotalCharges</th>\n",
       "      <th>Churn</th>\n",
       "      <th>TotalDays</th>\n",
       "      <th>Gender</th>\n",
       "      <th>SeniorCitizen</th>\n",
       "      <th>Partner</th>\n",
       "      <th>Dependents</th>\n",
       "      <th>InternetService</th>\n",
       "      <th>OnlineSecurity</th>\n",
       "      <th>OnlineBackup</th>\n",
       "      <th>DeviceProtection</th>\n",
       "      <th>TechSupport</th>\n",
       "      <th>StreamingTV</th>\n",
       "      <th>StreamingMovies</th>\n",
       "      <th>MultipleLines</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CustomerID</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>7590-VHVEG</td>\n",
       "      <td>2020-01-01</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>Month-to-month</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Electronic check</td>\n",
       "      <td>29.85</td>\n",
       "      <td>29.85</td>\n",
       "      <td>0</td>\n",
       "      <td>31</td>\n",
       "      <td>Female</td>\n",
       "      <td>0</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>DSL</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5575-GNVDE</td>\n",
       "      <td>2017-04-01</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>One year</td>\n",
       "      <td>No</td>\n",
       "      <td>Mailed check</td>\n",
       "      <td>56.95</td>\n",
       "      <td>1889.50</td>\n",
       "      <td>0</td>\n",
       "      <td>1036</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>DSL</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3668-QPYBK</td>\n",
       "      <td>2019-10-01</td>\n",
       "      <td>2019-12-01</td>\n",
       "      <td>Month-to-month</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Mailed check</td>\n",
       "      <td>53.85</td>\n",
       "      <td>108.15</td>\n",
       "      <td>1</td>\n",
       "      <td>61</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>DSL</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7795-CFOCW</td>\n",
       "      <td>2016-05-01</td>\n",
       "      <td>2020-02-01</td>\n",
       "      <td>One year</td>\n",
       "      <td>No</td>\n",
       "      <td>Bank transfer (automatic)</td>\n",
       "      <td>42.30</td>\n",
       "      <td>1840.75</td>\n",
       "      <td>0</td>\n",
       "      <td>1371</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>DSL</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9237-HQITU</td>\n",
       "      <td>2019-09-01</td>\n",
       "      <td>2019-11-01</td>\n",
       "      <td>Month-to-month</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Electronic check</td>\n",
       "      <td>70.70</td>\n",
       "      <td>151.65</td>\n",
       "      <td>1</td>\n",
       "      <td>61</td>\n",
       "      <td>Female</td>\n",
       "      <td>0</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Fiber optic</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            BeginDate    EndDate            Type PaperlessBilling  \\\n",
       "CustomerID                                                          \n",
       "7590-VHVEG 2020-01-01 2020-02-01  Month-to-month              Yes   \n",
       "5575-GNVDE 2017-04-01 2020-02-01        One year               No   \n",
       "3668-QPYBK 2019-10-01 2019-12-01  Month-to-month              Yes   \n",
       "7795-CFOCW 2016-05-01 2020-02-01        One year               No   \n",
       "9237-HQITU 2019-09-01 2019-11-01  Month-to-month              Yes   \n",
       "\n",
       "                        PaymentMethod  MonthlyCharges  TotalCharges  Churn  \\\n",
       "CustomerID                                                                   \n",
       "7590-VHVEG           Electronic check           29.85         29.85      0   \n",
       "5575-GNVDE               Mailed check           56.95       1889.50      0   \n",
       "3668-QPYBK               Mailed check           53.85        108.15      1   \n",
       "7795-CFOCW  Bank transfer (automatic)           42.30       1840.75      0   \n",
       "9237-HQITU           Electronic check           70.70        151.65      1   \n",
       "\n",
       "            TotalDays  Gender  SeniorCitizen Partner Dependents  \\\n",
       "CustomerID                                                        \n",
       "7590-VHVEG         31  Female              0     Yes         No   \n",
       "5575-GNVDE       1036    Male              0      No         No   \n",
       "3668-QPYBK         61    Male              0      No         No   \n",
       "7795-CFOCW       1371    Male              0      No         No   \n",
       "9237-HQITU         61  Female              0      No         No   \n",
       "\n",
       "           InternetService OnlineSecurity OnlineBackup DeviceProtection  \\\n",
       "CustomerID                                                                \n",
       "7590-VHVEG             DSL             No          Yes               No   \n",
       "5575-GNVDE             DSL            Yes           No              Yes   \n",
       "3668-QPYBK             DSL            Yes          Yes               No   \n",
       "7795-CFOCW             DSL            Yes           No              Yes   \n",
       "9237-HQITU     Fiber optic             No           No               No   \n",
       "\n",
       "           TechSupport StreamingTV StreamingMovies MultipleLines  \n",
       "CustomerID                                                        \n",
       "7590-VHVEG          No          No              No           NaN  \n",
       "5575-GNVDE          No          No              No            No  \n",
       "3668-QPYBK          No          No              No            No  \n",
       "7795-CFOCW         Yes          No              No           NaN  \n",
       "9237-HQITU          No          No              No            No  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MonthlyCharges</th>\n",
       "      <th>TotalCharges</th>\n",
       "      <th>Churn</th>\n",
       "      <th>TotalDays</th>\n",
       "      <th>SeniorCitizen</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>count</td>\n",
       "      <td>7043.000000</td>\n",
       "      <td>7043.000000</td>\n",
       "      <td>7043.000000</td>\n",
       "      <td>7043.000000</td>\n",
       "      <td>7043.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>mean</td>\n",
       "      <td>64.761692</td>\n",
       "      <td>2279.734304</td>\n",
       "      <td>0.265370</td>\n",
       "      <td>985.996166</td>\n",
       "      <td>0.162147</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>std</td>\n",
       "      <td>30.090047</td>\n",
       "      <td>2266.794470</td>\n",
       "      <td>0.441561</td>\n",
       "      <td>747.587844</td>\n",
       "      <td>0.368612</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>min</td>\n",
       "      <td>18.250000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>25%</td>\n",
       "      <td>35.500000</td>\n",
       "      <td>398.550000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>273.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>50%</td>\n",
       "      <td>70.350000</td>\n",
       "      <td>1394.550000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>883.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>75%</td>\n",
       "      <td>89.850000</td>\n",
       "      <td>3786.600000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1676.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>max</td>\n",
       "      <td>118.750000</td>\n",
       "      <td>8684.800000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>2191.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       MonthlyCharges  TotalCharges        Churn    TotalDays  SeniorCitizen\n",
       "count     7043.000000   7043.000000  7043.000000  7043.000000    7043.000000\n",
       "mean        64.761692   2279.734304     0.265370   985.996166       0.162147\n",
       "std         30.090047   2266.794470     0.441561   747.587844       0.368612\n",
       "min         18.250000      0.000000     0.000000     0.000000       0.000000\n",
       "25%         35.500000    398.550000     0.000000   273.000000       0.000000\n",
       "50%         70.350000   1394.550000     0.000000   883.000000       0.000000\n",
       "75%         89.850000   3786.600000     1.000000  1676.000000       0.000000\n",
       "max        118.750000   8684.800000     1.000000  2191.000000       1.000000"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 7043 entries, 7590-VHVEG to 3186-AJIEK\n",
      "Data columns (total 21 columns):\n",
      "BeginDate           7043 non-null datetime64[ns]\n",
      "EndDate             7043 non-null datetime64[ns]\n",
      "Type                7043 non-null object\n",
      "PaperlessBilling    7043 non-null object\n",
      "PaymentMethod       7043 non-null object\n",
      "MonthlyCharges      7043 non-null float64\n",
      "TotalCharges        7043 non-null float64\n",
      "Churn               7043 non-null int64\n",
      "TotalDays           7043 non-null int64\n",
      "Gender              7043 non-null object\n",
      "SeniorCitizen       7043 non-null int64\n",
      "Partner             7043 non-null object\n",
      "Dependents          7043 non-null object\n",
      "InternetService     5517 non-null object\n",
      "OnlineSecurity      5517 non-null object\n",
      "OnlineBackup        5517 non-null object\n",
      "DeviceProtection    5517 non-null object\n",
      "TechSupport         5517 non-null object\n",
      "StreamingTV         5517 non-null object\n",
      "StreamingMovies     5517 non-null object\n",
      "MultipleLines       6361 non-null object\n",
      "dtypes: datetime64[ns](2), float64(2), int64(3), object(14)\n",
      "memory usage: 1.2+ MB\n"
     ]
    }
   ],
   "source": [
    "analyze_df(churn_df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Пропуски в `InternetService`, `OnlineSecurity`, `OnlineBackup`, `DeviceProtection`, `TechSupport`, `StreamingTV`, `StreamingMovies` и `MultipleLines` скорее всего означают, что услуги не подключены. \n",
    "Заменим пропуски на \"No\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "for column in ('InternetService', 'OnlineSecurity', 'OnlineBackup', 'DeviceProtection', 'TechSupport', 'StreamingTV', 'StreamingMovies', 'MultipleLines'):\n",
    "    churn_df[column] = churn_df[column].fillna('No')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div style=\"background: #cceeaa; padding: 5px; border: 1px solid green; border-radius: 5px;\">\n",
    "    <font color='green'> <b><u>КОММЕНТАРИЙ РЕВЬЮЕРА</u></b>\n",
    "</font>\n",
    "<font color='green'><br>согласен."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 7043 entries, 7590-VHVEG to 3186-AJIEK\n",
      "Data columns (total 21 columns):\n",
      "BeginDate           7043 non-null datetime64[ns]\n",
      "EndDate             7043 non-null datetime64[ns]\n",
      "Type                7043 non-null object\n",
      "PaperlessBilling    7043 non-null object\n",
      "PaymentMethod       7043 non-null object\n",
      "MonthlyCharges      7043 non-null float64\n",
      "TotalCharges        7043 non-null float64\n",
      "Churn               7043 non-null int64\n",
      "TotalDays           7043 non-null int64\n",
      "Gender              7043 non-null object\n",
      "SeniorCitizen       7043 non-null int64\n",
      "Partner             7043 non-null object\n",
      "Dependents          7043 non-null object\n",
      "InternetService     7043 non-null object\n",
      "OnlineSecurity      7043 non-null object\n",
      "OnlineBackup        7043 non-null object\n",
      "DeviceProtection    7043 non-null object\n",
      "TechSupport         7043 non-null object\n",
      "StreamingTV         7043 non-null object\n",
      "StreamingMovies     7043 non-null object\n",
      "MultipleLines       7043 non-null object\n",
      "dtypes: datetime64[ns](2), float64(2), int64(3), object(14)\n",
      "memory usage: 1.2+ MB\n"
     ]
    }
   ],
   "source": [
    "churn_df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "От пропусков избавились, теперь можно проанализировать данные"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div style=\"background: #cceeaa; padding: 5px; border: 1px solid green; border-radius: 5px;\">\n",
    "    <font color='green'> <b><u>КОММЕНТАРИЙ РЕВЬЮЕРА</u></b>\n",
    "</font>\n",
    "<font color='green'><br>Хороший блок получился."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"exploratory_data_analysis\"></a>\n",
    "## Исследовательский анализ данных"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Посмотрим распределение значений для категориальных данных "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x2520 with 15 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cols = 2\n",
    "rows = int(np.ceil(len(categorial_columns) / cols))\n",
    "\n",
    "plt.figure(figsize=(20, 35))\n",
    "\n",
    "for idx, column in enumerate(categorial_columns):\n",
    "    ax = plt.subplot(rows, cols, idx + 1)\n",
    "    \n",
    "    churn_df[column].value_counts().plot(\n",
    "        kind='bar', \n",
    "        title=column, \n",
    "        ax=ax,\n",
    "        rot=0,\n",
    "    )\n",
    "        \n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Выводы**:\n",
    "* Самый популярный способ оплаты - помесячный \n",
    "* Люди чаще предпочитают получать счета в электронном виде и оплачивать их при помощи электронного чека\n",
    "* Самый частый способ подключения - оптоволокно\n",
    "* Дополнительные услуги `OnlineSecurity`, `OnlineBackup`, `DeviceProtection`, `TechSupport`, `StreamingTV`, `StreamingMovies`, `MultipleLines` не пользуются популярностью у пользователей"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div style=\"background: #cceeaa; padding: 5px; border: 1px solid green; border-radius: 5px;\">\n",
    "    <font color='green'> <b><u>КОММЕНТАРИЙ РЕВЬЮЕРА</u></b>\n",
    "</font>\n",
    "<font color='green'><br>Познавательно. ОК."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Посмотрим распределение значений и диаграммы размаха для количественных данных "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MonthlyCharges</th>\n",
       "      <th>TotalCharges</th>\n",
       "      <th>TotalDays</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>count</td>\n",
       "      <td>7043.000000</td>\n",
       "      <td>7043.000000</td>\n",
       "      <td>7043.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>mean</td>\n",
       "      <td>64.761692</td>\n",
       "      <td>2279.734304</td>\n",
       "      <td>985.996166</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>std</td>\n",
       "      <td>30.090047</td>\n",
       "      <td>2266.794470</td>\n",
       "      <td>747.587844</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>min</td>\n",
       "      <td>18.250000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>25%</td>\n",
       "      <td>35.500000</td>\n",
       "      <td>398.550000</td>\n",
       "      <td>273.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>50%</td>\n",
       "      <td>70.350000</td>\n",
       "      <td>1394.550000</td>\n",
       "      <td>883.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>75%</td>\n",
       "      <td>89.850000</td>\n",
       "      <td>3786.600000</td>\n",
       "      <td>1676.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>max</td>\n",
       "      <td>118.750000</td>\n",
       "      <td>8684.800000</td>\n",
       "      <td>2191.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       MonthlyCharges  TotalCharges    TotalDays\n",
       "count     7043.000000   7043.000000  7043.000000\n",
       "mean        64.761692   2279.734304   985.996166\n",
       "std         30.090047   2266.794470   747.587844\n",
       "min         18.250000      0.000000     0.000000\n",
       "25%         35.500000    398.550000   273.000000\n",
       "50%         70.350000   1394.550000   883.000000\n",
       "75%         89.850000   3786.600000  1676.000000\n",
       "max        118.750000   8684.800000  2191.000000"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_numerical(churn_df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Выводы**:\n",
    "* Значения в столбце `TotalCharges` убывают, в то же время в столбце `MonthlyCharges` виден пик в райне 85 - это значит, что пользователи, которые платят за тариф много на долго не остаются \n",
    "* Также у `MonthlyCharges` виден пик в районе 20, что является минимальным значением\n",
    "* В столбце `TotalDays` видно 2 пика - в районе новых клиентов и лояльных, которые долго пользуются услугами \n",
    "* Аномальных выбросов не видно"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div style=\"background: #cceeaa; padding: 5px; border: 1px solid green; border-radius: 5px;\">\n",
    "    <font color='green'> <b><u>КОММЕНТАРИЙ РЕВЬЮЕРА</u></b>\n",
    "</font>\n",
    "<font color='green'><br>ещё, гистограмма показывает бимодальность MonthlyCharges - несколько горбов) Это говорит о том, что в самой выборке есть несколько однородных выборок. Здесь, может быть, это завязано на количество услуг, или \"болатый/бедный\" клиент. В общем - можно попробовать выделить ещё фактор, который как раз объясняет эту бимодальность."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Посмотрим на распределение категорильных признаков относительно целевого признака `Result`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x2160 with 15 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cols = 2\n",
    "rows = int(np.ceil(len(categorial_columns) / cols))\n",
    "\n",
    "plt.figure(figsize=(20, 30))\n",
    "\n",
    "for idx, column in enumerate(categorial_columns):\n",
    "    churn_grouped_df = churn_df.groupby([column, 'Churn'])['Churn'].count()\n",
    "    churn_grouped_df = churn_grouped_df.unstack()\n",
    "    \n",
    "    ax = plt.subplot(rows, cols, idx + 1)\n",
    "\n",
    "    ax.bar(churn_grouped_df.T.index, churn_grouped_df.T[churn_grouped_df.T.columns[0]], label=churn_grouped_df.T.columns[0])\n",
    "    bottom = np.array(churn_grouped_df.T[churn_grouped_df.T.columns[0]])\n",
    "    for color_idx, value in enumerate(churn_grouped_df.T.columns[1:]): \n",
    "        ax.bar(churn_grouped_df.T.index, churn_grouped_df.T[value], bottom=bottom, label=value)\n",
    "        bottom += np.array(churn_grouped_df.T[value])\n",
    "    plt.title(column)\n",
    "    plt.legend()\n",
    "        \n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Выводы**:\n",
    "* Доля отказов среди тех, кто использует помесячную оплату выше, чем среди тех, кто сразу платит за 1/2 года \n",
    "* Дополнительные услуги слабо влияют на отказы\n",
    "* Пользователи использующие \"Electronic check\" для оплаты чаще отказываются от услуг, возможно в оплате онлайн есть какие-то неудобства\n",
    "* Пенсионеры уходят чаще, но скорее всего не из-за качества услуг \n",
    "* Если пользователь не пользуется интернетом, то вероятность отказа ниже. Скорее всего это связано с тем, что более возрастные люди не хотят разбираться в том, как сменить поставщика услуг"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div style=\"background: #cceeaa; padding: 5px; border: 1px solid green; border-radius: 5px;\">\n",
    "    <font color='green'> <b><u>КОММЕНТАРИЙ РЕВЬЮЕРА</u></b>\n",
    "</font>\n",
    "<font color='green'><br>классно"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Посмотрим на распределение количественных признаков относительно целевого признака `Result`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MonthlyCharges</th>\n",
       "      <th>TotalCharges</th>\n",
       "      <th>TotalDays</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>count</td>\n",
       "      <td>1869.000000</td>\n",
       "      <td>1869.000000</td>\n",
       "      <td>1869.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>mean</td>\n",
       "      <td>74.441332</td>\n",
       "      <td>1531.796094</td>\n",
       "      <td>547.352060</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>std</td>\n",
       "      <td>24.666053</td>\n",
       "      <td>1890.822994</td>\n",
       "      <td>594.389607</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>min</td>\n",
       "      <td>18.850000</td>\n",
       "      <td>18.850000</td>\n",
       "      <td>30.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>25%</td>\n",
       "      <td>56.150000</td>\n",
       "      <td>134.500000</td>\n",
       "      <td>61.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>50%</td>\n",
       "      <td>79.650000</td>\n",
       "      <td>703.550000</td>\n",
       "      <td>304.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>75%</td>\n",
       "      <td>94.200000</td>\n",
       "      <td>2331.300000</td>\n",
       "      <td>883.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>max</td>\n",
       "      <td>118.350000</td>\n",
       "      <td>8684.800000</td>\n",
       "      <td>2191.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       MonthlyCharges  TotalCharges    TotalDays\n",
       "count     1869.000000   1869.000000  1869.000000\n",
       "mean        74.441332   1531.796094   547.352060\n",
       "std         24.666053   1890.822994   594.389607\n",
       "min         18.850000     18.850000    30.000000\n",
       "25%         56.150000    134.500000    61.000000\n",
       "50%         79.650000    703.550000   304.000000\n",
       "75%         94.200000   2331.300000   883.000000\n",
       "max        118.350000   8684.800000  2191.000000"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_numerical(churn_df[churn_df['Churn'] == 1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MonthlyCharges</th>\n",
       "      <th>TotalCharges</th>\n",
       "      <th>TotalDays</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>count</td>\n",
       "      <td>5174.000000</td>\n",
       "      <td>5174.000000</td>\n",
       "      <td>5174.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>mean</td>\n",
       "      <td>61.265124</td>\n",
       "      <td>2549.911442</td>\n",
       "      <td>1144.447236</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>std</td>\n",
       "      <td>31.092648</td>\n",
       "      <td>2329.954215</td>\n",
       "      <td>733.897937</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>min</td>\n",
       "      <td>18.250000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>25%</td>\n",
       "      <td>25.100000</td>\n",
       "      <td>572.900000</td>\n",
       "      <td>457.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>50%</td>\n",
       "      <td>64.425000</td>\n",
       "      <td>1679.525000</td>\n",
       "      <td>1157.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>75%</td>\n",
       "      <td>88.400000</td>\n",
       "      <td>4262.850000</td>\n",
       "      <td>1857.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>max</td>\n",
       "      <td>118.750000</td>\n",
       "      <td>8672.450000</td>\n",
       "      <td>2191.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       MonthlyCharges  TotalCharges    TotalDays\n",
       "count     5174.000000   5174.000000  5174.000000\n",
       "mean        61.265124   2549.911442  1144.447236\n",
       "std         31.092648   2329.954215   733.897937\n",
       "min         18.250000      0.000000     0.000000\n",
       "25%         25.100000    572.900000   457.000000\n",
       "50%         64.425000   1679.525000  1157.000000\n",
       "75%         88.400000   4262.850000  1857.000000\n",
       "max        118.750000   8672.450000  2191.000000"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_numerical(churn_df[churn_df['Churn'] == 0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Выводы**:\n",
    "* Со временем вероятность отказа от услуг снижается \n",
    "* Обычно отказывались от услуг те, кто платил больше\n",
    "* Текущие клиенты платят немного меньше в месяц, чем те, кто отказался "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Рассмотррим матрицу корреляции "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Churn             1.000000\n",
       "MonthlyCharges    0.193356\n",
       "SeniorCitizen     0.150889\n",
       "TotalCharges     -0.198324\n",
       "TotalDays        -0.352673\n",
       "Name: Churn, dtype: float64"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "churn_df.corr()['Churn'].sort_values(ascending=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Наши предположения подтвердились: \n",
    "* чем дольше клиент пользовался услугами, тем меньше вероятность, что он уйдет\n",
    "* чем выше цена за услуги в месяц, тем больше вероятность, что клиент уйдет"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div style=\"background: #cceeaa; padding: 5px; border: 1px solid green; border-radius: 5px;\">\n",
    "    <font color='green'> <b><u>КОММЕНТАРИЙ РЕВЬЮЕРА</u></b>\n",
    "</font>\n",
    "<font color='green'><br>\n",
    "Отдичный анализ. Напомню только пару моментов:<br>\n",
    "1) графический анализ - это предположение/гипотеза. И чтобы её подтвердить или опровергнуть также следует применять аналитические методы анализа из области \"проверка гипотез\".<br>\n",
    "2) Техника корреляций - ТОЛЬКО для количественных данных. бинарные и номинальные - лучше не смотреть."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-success\">\n",
    "Выведем корреляцию между всеми количественными данными\n",
    "</div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MonthlyCharges</th>\n",
       "      <th>TotalCharges</th>\n",
       "      <th>TotalDays</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>MonthlyCharges</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.651174</td>\n",
       "      <td>0.247754</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>TotalCharges</td>\n",
       "      <td>0.651174</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.826109</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>TotalDays</td>\n",
       "      <td>0.247754</td>\n",
       "      <td>0.826109</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                MonthlyCharges  TotalCharges  TotalDays\n",
       "MonthlyCharges        1.000000      0.651174   0.247754\n",
       "TotalCharges          0.651174      1.000000   0.826109\n",
       "TotalDays             0.247754      0.826109   1.000000"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "churn_df[numeric_columns].corr()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Можно заметить, что у нас присутствует проблема мультиколлинеарности в данных - корреляция между `TotalCharges` и `TotalDays` - 0.82, что логично - чем дольше клиент платит, тем больше итоговая сумма. Удалим поле `TotalCharges` "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "del churn_df['TotalCharges']\n",
    "numeric_columns = ['MonthlyCharges', 'TotalDays']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"prepare_data_for_models\"></a>\n",
    "## Подготовка данных для моделей"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Удалим ненужные для обучения столбцы `BeginDate` и `EndDate`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "del churn_df['BeginDate'], churn_df['EndDate']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div style=\"background: #cceeaa; padding: 5px; border: 1px solid green; border-radius: 5px;\">\n",
    "    <font color='green'> <b><u>КОММЕНТАРИЙ РЕВЬЮЕРА</u></b>\n",
    "</font>\n",
    "<font color='green'><br>\n",
    "ОТЛИЧНО, что удаляешь факторы-даты!!!!! Если у нас тема НЕ временные ряды, то такие факторы лучше не использовать. Они \"нагонят\" метрики, но бизнес-эффективность модели  - рухнет."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Применим технику OHE (One-Hot Encoding) для того, чтобы преобразовать категоральные признаки в численные "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "churn_df = pd.get_dummies(churn_df, drop_first=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MonthlyCharges</th>\n",
       "      <th>Churn</th>\n",
       "      <th>TotalDays</th>\n",
       "      <th>SeniorCitizen</th>\n",
       "      <th>Type_One year</th>\n",
       "      <th>Type_Two year</th>\n",
       "      <th>PaperlessBilling_Yes</th>\n",
       "      <th>PaymentMethod_Credit card (automatic)</th>\n",
       "      <th>PaymentMethod_Electronic check</th>\n",
       "      <th>PaymentMethod_Mailed check</th>\n",
       "      <th>Gender_Male</th>\n",
       "      <th>Partner_Yes</th>\n",
       "      <th>Dependents_Yes</th>\n",
       "      <th>InternetService_Fiber optic</th>\n",
       "      <th>InternetService_No</th>\n",
       "      <th>OnlineSecurity_Yes</th>\n",
       "      <th>OnlineBackup_Yes</th>\n",
       "      <th>DeviceProtection_Yes</th>\n",
       "      <th>TechSupport_Yes</th>\n",
       "      <th>StreamingTV_Yes</th>\n",
       "      <th>StreamingMovies_Yes</th>\n",
       "      <th>MultipleLines_Yes</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CustomerID</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>7590-VHVEG</td>\n",
       "      <td>29.85</td>\n",
       "      <td>0</td>\n",
       "      <td>31</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5575-GNVDE</td>\n",
       "      <td>56.95</td>\n",
       "      <td>0</td>\n",
       "      <td>1036</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3668-QPYBK</td>\n",
       "      <td>53.85</td>\n",
       "      <td>1</td>\n",
       "      <td>61</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7795-CFOCW</td>\n",
       "      <td>42.30</td>\n",
       "      <td>0</td>\n",
       "      <td>1371</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9237-HQITU</td>\n",
       "      <td>70.70</td>\n",
       "      <td>1</td>\n",
       "      <td>61</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            MonthlyCharges  Churn  TotalDays  SeniorCitizen  Type_One year  \\\n",
       "CustomerID                                                                   \n",
       "7590-VHVEG           29.85      0         31              0              0   \n",
       "5575-GNVDE           56.95      0       1036              0              1   \n",
       "3668-QPYBK           53.85      1         61              0              0   \n",
       "7795-CFOCW           42.30      0       1371              0              1   \n",
       "9237-HQITU           70.70      1         61              0              0   \n",
       "\n",
       "            Type_Two year  PaperlessBilling_Yes  \\\n",
       "CustomerID                                        \n",
       "7590-VHVEG              0                     1   \n",
       "5575-GNVDE              0                     0   \n",
       "3668-QPYBK              0                     1   \n",
       "7795-CFOCW              0                     0   \n",
       "9237-HQITU              0                     1   \n",
       "\n",
       "            PaymentMethod_Credit card (automatic)  \\\n",
       "CustomerID                                          \n",
       "7590-VHVEG                                      0   \n",
       "5575-GNVDE                                      0   \n",
       "3668-QPYBK                                      0   \n",
       "7795-CFOCW                                      0   \n",
       "9237-HQITU                                      0   \n",
       "\n",
       "            PaymentMethod_Electronic check  PaymentMethod_Mailed check  \\\n",
       "CustomerID                                                               \n",
       "7590-VHVEG                               1                           0   \n",
       "5575-GNVDE                               0                           1   \n",
       "3668-QPYBK                               0                           1   \n",
       "7795-CFOCW                               0                           0   \n",
       "9237-HQITU                               1                           0   \n",
       "\n",
       "            Gender_Male  Partner_Yes  Dependents_Yes  \\\n",
       "CustomerID                                             \n",
       "7590-VHVEG            0            1               0   \n",
       "5575-GNVDE            1            0               0   \n",
       "3668-QPYBK            1            0               0   \n",
       "7795-CFOCW            1            0               0   \n",
       "9237-HQITU            0            0               0   \n",
       "\n",
       "            InternetService_Fiber optic  InternetService_No  \\\n",
       "CustomerID                                                    \n",
       "7590-VHVEG                            0                   0   \n",
       "5575-GNVDE                            0                   0   \n",
       "3668-QPYBK                            0                   0   \n",
       "7795-CFOCW                            0                   0   \n",
       "9237-HQITU                            1                   0   \n",
       "\n",
       "            OnlineSecurity_Yes  OnlineBackup_Yes  DeviceProtection_Yes  \\\n",
       "CustomerID                                                               \n",
       "7590-VHVEG                   0                 1                     0   \n",
       "5575-GNVDE                   1                 0                     1   \n",
       "3668-QPYBK                   1                 1                     0   \n",
       "7795-CFOCW                   1                 0                     1   \n",
       "9237-HQITU                   0                 0                     0   \n",
       "\n",
       "            TechSupport_Yes  StreamingTV_Yes  StreamingMovies_Yes  \\\n",
       "CustomerID                                                          \n",
       "7590-VHVEG                0                0                    0   \n",
       "5575-GNVDE                0                0                    0   \n",
       "3668-QPYBK                0                0                    0   \n",
       "7795-CFOCW                1                0                    0   \n",
       "9237-HQITU                0                0                    0   \n",
       "\n",
       "            MultipleLines_Yes  \n",
       "CustomerID                     \n",
       "7590-VHVEG                  0  \n",
       "5575-GNVDE                  0  \n",
       "3668-QPYBK                  0  \n",
       "7795-CFOCW                  0  \n",
       "9237-HQITU                  0  "
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "churn_df.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div style=\"background: #cceeaa; padding: 5px; border: 1px solid green; border-radius: 5px;\">\n",
    "    <font color='green'> <b><u>КОММЕНТАРИЙ РЕВЬЮЕРА</u></b>\n",
    "</font>\n",
    "<font color='green'><br>\n",
    "принято."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Выделим фичи и целевой показатель `Churn` "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "target = churn_df['Churn'] \n",
    "features = churn_df.drop(columns=['Churn'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Разделим датафрейм на обучающую и тестовую выбороки"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "features_train, features_test, target_train, target_test = train_test_split(features, target, test_size=0.2, random_state=RANDOM_STATE)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Проведем масштабирование количественных признаков"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = StandardScaler()\n",
    "\n",
    "scaler.fit(features_train[numeric_columns])\n",
    "\n",
    "features_train = features_train.copy()\n",
    "features_train[numeric_columns] = scaler.transform(features_train[numeric_columns])\n",
    "\n",
    "features_test = features_test.copy()\n",
    "features_test[numeric_columns] = scaler.transform(features_test[numeric_columns])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Проверим баланс классов"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "target_train.value_counts().plot(\n",
    "    kind='bar', \n",
    "    figsize=(10, 5), \n",
    "    legend=False, \n",
    "    title='Распределение значений целевого признака', \n",
    ").set(xlabel='Значение', ylabel='Кол-во значений')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "ratio = target_train[target_train == 1].count() / target_train[target_train == 0].count()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Видно, что присутсвует дисбаланс классов, надо будет провести upsampling, т.е увеличить число объектов меньшего класса путём их многократного копирования."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "features_train, target_train = upsample(features_train, target_train, 3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "target_train.value_counts().plot(\n",
    "    kind='bar', \n",
    "    figsize=(10, 5), \n",
    "    legend=False, \n",
    "    title='Распределение значений целевого признака', \n",
    ").set(xlabel='Значение', ylabel='Кол-во значений')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Видно, что распределение значений целевого признака выправилось"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div style=\"background: #cceeaa; padding: 5px; border: 1px solid green; border-radius: 5px;\">\n",
    "    <font color='green'> <b><u>КОММЕНТАРИЙ РЕВЬЮЕРА</u></b>\n",
    "</font>\n",
    "<font color='green'><br>\n",
    "да, понял."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"models_train\"></a>\n",
    "## Обучение моделей"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### LogisticRegression"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Сначала попробуем обучить модель логистической регрессии"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'C': 100}\n"
     ]
    }
   ],
   "source": [
    "param_grid = {'C': [0.001, 0.01, 0.1, 1, 10, 100, 1000]}\n",
    "\n",
    "model = LogisticRegression(solver='lbfgs', n_jobs=-1)\n",
    "\n",
    "search = GridSearchCV(model, param_grid, cv=5, scoring='roc_auc')\n",
    "search.fit(features_train, target_train)\n",
    "\n",
    "print(search.best_params_)\n",
    "\n",
    "predicted_probalities = search.best_estimator_.predict_proba(features_test)[:, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROC-AUC score 0.8437636088838795\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_roc_curve(target_test, predicted_probalities)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### RandomForestClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'max_depth': 20, 'min_samples_leaf': 2, 'min_samples_split': 2, 'n_estimators': 100}\n"
     ]
    }
   ],
   "source": [
    "param_grid = {\n",
    "    'max_depth': [5, 10, 20],\n",
    "    'min_samples_split': [2, 10],\n",
    "    'min_samples_leaf': [2, 10],\n",
    "    'n_estimators': [10, 50, 100], \n",
    "}\n",
    "\n",
    "model = RandomForestClassifier(random_state=RANDOM_STATE)\n",
    "\n",
    "search = GridSearchCV(model, param_grid, cv=5, scoring='roc_auc')\n",
    "search.fit(features_train, target_train)\n",
    "\n",
    "print(search.best_params_)\n",
    "\n",
    "predicted_probalities = search.best_estimator_.predict_proba(features_test)[:, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROC-AUC score 0.8500256167225964\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_roc_curve(target_test, predicted_probalities)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### CatBoostClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'iterations': 100, 'l2_leaf_reg': 3, 'learning_rate': 0.3}\n"
     ]
    }
   ],
   "source": [
    "param_grid = {\n",
    "    'learning_rate': [0.1, 0.3],\n",
    "    'iterations': [50, 100],\n",
    "    'l2_leaf_reg': [3, 9]\n",
    "}\n",
    "\n",
    "model = CatBoostClassifier(random_state=RANDOM_STATE, verbose=0)\n",
    "\n",
    "search = GridSearchCV(model, param_grid, cv=5, scoring='roc_auc')\n",
    "search.fit(features_train, target_train)\n",
    "\n",
    "print(search.best_params_)\n",
    "\n",
    "predicted_probalities = search.best_estimator_.predict_proba(features_test)[:, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROC-AUC score 0.9142557061249584\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_roc_curve(target_test, predicted_probalities)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-success\">\n",
    "Catboost показал лучшее значение метрики ROC-AUC - 0.91 \n",
    "<br>\n",
    "Вычилим важность факторов\n",
    "</div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>importances</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>TotalDays</td>\n",
       "      <td>59.023932</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>MonthlyCharges</td>\n",
       "      <td>11.664119</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>InternetService_Fiber optic</td>\n",
       "      <td>4.060303</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Type_Two year</td>\n",
       "      <td>3.333899</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Type_One year</td>\n",
       "      <td>1.953484</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>PaymentMethod_Electronic check</td>\n",
       "      <td>1.937646</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>OnlineSecurity_Yes</td>\n",
       "      <td>1.789360</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>TechSupport_Yes</td>\n",
       "      <td>1.653227</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>StreamingMovies_Yes</td>\n",
       "      <td>1.563230</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Gender_Male</td>\n",
       "      <td>1.514801</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>MultipleLines_Yes</td>\n",
       "      <td>1.444294</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>SeniorCitizen</td>\n",
       "      <td>1.383704</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Partner_Yes</td>\n",
       "      <td>1.222827</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>StreamingTV_Yes</td>\n",
       "      <td>1.135844</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>PaymentMethod_Mailed check</td>\n",
       "      <td>1.103624</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>PaperlessBilling_Yes</td>\n",
       "      <td>1.091416</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>OnlineBackup_Yes</td>\n",
       "      <td>1.081362</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Dependents_Yes</td>\n",
       "      <td>0.981419</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>PaymentMethod_Credit card (automatic)</td>\n",
       "      <td>0.971449</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>DeviceProtection_Yes</td>\n",
       "      <td>0.644573</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>InternetService_No</td>\n",
       "      <td>0.445488</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                       importances\n",
       "TotalDays                                59.023932\n",
       "MonthlyCharges                           11.664119\n",
       "InternetService_Fiber optic               4.060303\n",
       "Type_Two year                             3.333899\n",
       "Type_One year                             1.953484\n",
       "PaymentMethod_Electronic check            1.937646\n",
       "OnlineSecurity_Yes                        1.789360\n",
       "TechSupport_Yes                           1.653227\n",
       "StreamingMovies_Yes                       1.563230\n",
       "Gender_Male                               1.514801\n",
       "MultipleLines_Yes                         1.444294\n",
       "SeniorCitizen                             1.383704\n",
       "Partner_Yes                               1.222827\n",
       "StreamingTV_Yes                           1.135844\n",
       "PaymentMethod_Mailed check                1.103624\n",
       "PaperlessBilling_Yes                      1.091416\n",
       "OnlineBackup_Yes                          1.081362\n",
       "Dependents_Yes                            0.981419\n",
       "PaymentMethod_Credit card (automatic)     0.971449\n",
       "DeviceProtection_Yes                      0.644573\n",
       "InternetService_No                        0.445488"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "importances_df = pd.DataFrame(search.best_estimator_.feature_importances_, index = features_test.columns, columns=['importances'])\n",
    "importances_df = importances_df.sort_values(by='importances', ascending=False)\n",
    "importances_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlAAAAIACAYAAABNbFKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAgAElEQVR4nOzdd7hcVbnH8e8vCRBaAoEYg5QAQhApAZKIgIog5YrSRIqohF4EFbwoKtdrFwUvIKKCgBRBKUoRlE4ERIFQAiJIiUGCCJEapSXhvX+sNcnk5Jycs89Ze84k/D7Pkydn9szsd09/91rvWksRgZmZmZn13ID+PgAzMzOzhY0TKDMzM7OKnECZmZmZVeQEyszMzKwiJ1BmZmZmFTmBMjMzM6vICZSZmZlZRU6gzBYRkqZKekXSvyU9L+kqSav093H1J0mjJIWkQf19LGa2aHECZbZo+XBELAOMBJ4GTunn4zEzWyQ5gTJbBEXEq8AlwLqNbZJ2kHSPpJckPSHpq03XvT1v2zxfniDp1vz3kpJuk/TZfHlLSdOa40m6VdKE/PcAScdKelzSM5LOlTS06bZb5P29kGNOkLRHbjn7t6TZkl5tXM73+aqkn/fksUtaXtLVkp4BDs+bvy5puqQrJA1put2Vefvz+e+Vm/YzUdIB+e8187HuKGmlpmN9XdLMpsvvybc/UNKjkp7LMVdq2m9I+rSkKZL+Jel4SQO6e+6aWtMasf4sacuePCdmVp4TKLNFkKSlgD2APzVt/g/wSWA5YAfgUEk7A0TEo8DuwC8lrdW0nwHAz4E7IuKkHoafkP+9H1gDWAb4Yd7fasDvSC1jw4ExwL0RcWFELJNbz24BDm+6XNU3gBeB1YBX87angFWB2cBX87YBwM/y7VYFXmkcZzNJI4FrgC9FxBUR8Y+mY/s2MOfYI+IWSVsB3yE9nyOBx4FfdtjtLsBYYGNgJ2C/vH0CXTx3TZYDlgUuAk7o6ZNiZmU5gTJbtFwm6QVSArENcHzjioiYGBH3R8QbEXEf8AvgfU3X/xH4CinBGZ43nwCsCBxV4Rj2Bv4vIqZExL+BLwJ75jqkjwHXR8QvImJmRDwbEff2+tF27sPAqRHxCnBG3vbjfPlkYFeAHPtXEfFyRMwAvkXT85EtT0qezo+I83oYf2/grIi4OyJeIz3+d0sa1XSb70bEcxHxd+AkYK+m+3b13DUTMBB4tofHZGaFOYEyW7TsHBHLAYNJ3Ve/l/RWAEnvknRT7rJ6ETiElBw124b0o/x1YBNga2A08JYOt1spd8G9kBO2TZuvI7W6NDwODAJGAKsAj/Xyse2e4/1L0nWS1ujidiOA6V1c9wzQeD6WknRa7i57CbgZWE7SwKbbfw34N7CVJPXwOOd5/DkRehZ4W9Ntnmj6+/F8n/nuy7zPXcO/8jF9FvhuD4/JzApzAmW2CIqI2RHxa1KX1RZ58wXAFcAqETEU+AmpJQMASduQupXeB5wHvEBKqH5CaiVp9o+IWK7xj3m7Cv9B6hZrWBWYRSpqfwJYs5cP66IcayXg76Tus85MZ/7EsOEt+TgAPkdKDt8VEUOA9+btzYnSRaTnT8ARPTzOeR6/pKWBFYAnm27TPDpy1Xyf+e7LvM9dw4oRsRSp6+9Xkpbs4XGZWUFOoMwWQUp2InVBPZg3Lws8FxGvShpP6k5r3H4w8CPg0FyAfhvwWEQ8Q6rnGSNp+x6G/wVwpKTVJTXXCc0Czgc+IGl3SYMkrSBpTJXHFhGvk1pguvr++i1wWE4sDsjbDs2XPw38Jm9bllT39IKkYcD/drKvWyPiDVKN0lcW0OrV7BfAvpLGSFqC9Phvj4ipTbc5OhexrwJ8Briw6b5dPXcdzQaGAov34JjMrDAnUGaLlt/kkWsvkWp69omIB/J1h5FGo80g1Tpd1HS/Y4E/RcQNHXeY63gOAU7tYWvHWaQWrJuBv5EKuY/I+/o78EFS689zwL3Ahj18bLtImibpSVLx9bFd3O5YUg3X48ASedtbSa1fg4H/ydtOApYkdYn9Cbi6q8AR8TBwHHBGd115EXF9jvErUvH6msCeHW52OXAX6fFfBZyZt3f53DV5Ib/G5wIHR8SLCzoeM6uHIqK/j8HMrBa5cPtvwGJdtOK0nKQA1sojH81sIeUWKDMzM7OKnECZmZmZVeQuPDMzM7OK3AJlZmZmVpETKDMzM7OKOi4PUKsVV1wxRo0a1cqQZmZmZr1y1113/Ssihnd2XUsTqFGjRjFp0qRWhjQzMzPrFUmPd3Wdu/DMzMzMKnICZWZmZlaREygzMzOzinpUAyVpOeAMYD0gSAtr/pW0AOYoYCqwe0Q8X8tRmpmZLaJmzpzJtGnTePXVV/v7UN60Bg8ezMorr8xiiy3W4/v0tIj8ZODqiNhN0uLAUsCXgBsi4jhJxwDHAF+oetBmZmZvZtOmTWPZZZdl1KhRdLNWtdUgInj22WeZNm0aq6++eo/v120XnqShwHvJq4VHxOsR8QKwE3BOvtk5wM6Vj9rMzOxN7tVXX2WFFVZw8tRPJLHCCitUbgHsSQ3U6sB04GeS7pF0hqSlgRER8VS+zT+BEZUim5mZGYCTp37Wm+e/JwnUIGBj4McRsRHwH1J33RyRFtTrdFE9SQdJmiRp0vTp0ysfoJmZmdVrs802a2m8qVOncsEFF7Q0Zmk9qYGaBkyLiNvz5UtICdTTkkZGxFOSRgLPdHbniDgdOB1g7NixXrnYzMxsAUYdc1XR/U09bodub3PbbbcVjbkgs2bNmpNAfexjH2tZ3NK6bYGKiH8CT0ganTdtDfwFuALYJ2/bB7i8liM0MzOzWi2zzDIATJw4kfe9733stNNOrLHGGhxzzDGcf/75jB8/nvXXX5/HHnsMgAkTJnDIIYcwduxY1l57ba688kog1XPtu+++rL/++my00UbcdNNNAJx99tnsuOOObLXVVmy99dYcc8wx3HLLLYwZM4YTTzyRqVOn8p73vIeNN96YjTfeeE5CN3HiRLbcckt222031llnHfbee29SpxfceeedbLbZZmy44YaMHz+eGTNmMHv2bI4++mjGjRvHBhtswGmnnQbAU089xXvf+17GjBnDeuutxy233NLn56yno/COAM7PI/CmAPuSkq+LJO0PPA7s3uejMTMzs341efJkHnzwQYYNG8Yaa6zBAQccwB133MHJJ5/MKaecwkknnQSkbrg77riDxx57jPe///08+uijnHrqqUji/vvv56GHHmLbbbfl4YcfBuDuu+/mvvvuY9iwYUycOJETTjhhTuL18ssvc9111zF48GAeeeQR9tprrzlLv91zzz088MADrLTSSmy++eb84Q9/YPz48eyxxx5ceOGFjBs3jpdeeokll1ySM888k6FDh3LnnXfy2muvsfnmm7Ptttvy61//mu22244vf/nLzJ49m5dffrnPz1OPEqiIuBcY28lVW/f5COh9c2VPmiXNzMys58aNG8fIkSMBWHPNNdl2220BWH/99ee0KAHsvvvuDBgwgLXWWos11liDhx56iFtvvZUjjjgCgHXWWYfVVlttTgK1zTbbMGzYsE5jzpw5k8MPP5x7772XgQMHzrkPwPjx41l55ZUBGDNmDFOnTmXo0KGMHDmScePGATBkyBAArr32Wu677z4uueQSAF588UUeeeQRxo0bx3777cfMmTPZeeedGTNmTJ+fp5YuJmxmZmbtbYkllpjz94ABA+ZcHjBgALNmzZpzXceRa92NZFt66aW7vO7EE09kxIgRTJ48mTfeeIPBgwd3ejwDBw6c5xg6ighOOeUUtttuu/muu/nmm7nqqquYMGECRx11FJ/85CcXeLzd8VIuZmZmVtnFF1/MG2+8wWOPPcaUKVMYPXo073nPezj//PMBePjhh/n73//O6NGj57vvsssuy4wZM+ZcfvHFFxk5ciQDBgzgvPPOY/bs2QuMPXr0aJ566inuvPNOAGbMmMGsWbPYbrvt+PGPf8zMmTPnHMN//vMfHn/8cUaMGMGBBx7IAQccwN13393nx+8WKDMzM6ts1VVXZfz48bz00kv85Cc/YfDgwRx22GEceuihrL/++gwaNIizzz57nhakhg022ICBAwey4YYbMmHCBA477DA+8pGPcO6557L99tsvsLUKYPHFF+fCCy/kiCOO4JVXXmHJJZfk+uuv54ADDmDq1KlsvPHGRATDhw/nsssuY+LEiRx//PEstthiLLPMMpx77rl9fvxqVLO3wtixY6NRFNbMNVBmZvZm9eCDD/KOd7yjvw+jkgkTJvChD32I3Xbbrb8PpZjOXgdJd0VEZzXg7sIzMzMzq8pdeGZmZlbJ2Wef3d+H0O/cAmVmZmZWkRMoMzOzftbKemSbX2+efydQZmZm/Wjw4ME8++yzTqL6SUTw7LPPzjP3VE+4BsrMzKwfrbzyykybNo3p06f396G8aQ0ePHjObOc95QTKzMysHy222GKsvvrq/X0YVpG78MzMzMwqcgJlZmZmVpETKDMzM7OKnECZmZmZVeQEyszMzKwiJ1BmZmZmFTmBMjMzM6vICZSZmZlZRU6gzMzMzCpyAmVmZmZWkRMoMzMzs4qcQJmZmZlV5ATKzMzMrCInUGZmZmYVOYEyMzMzq8gJlJmZmVlFTqDMzMzMKnICZWZmZlaREygzMzOzipxAmZmZmVXkBMrMzMysIidQZmZmZhU5gTIzMzOryAmUmZmZWUWDenIjSVOBGcBsYFZEjJU0DLgQGAVMBXaPiOfrOUwzMzOz9lGlBer9ETEmIsbmy8cAN0TEWsAN+bKZmZnZIq8vXXg7Aefkv88Bdu774ZiZmZm1v54mUAFcK+kuSQflbSMi4qn89z+BEZ3dUdJBkiZJmjR9+vQ+Hq6ZmZlZ/+tRDRSwRUQ8KektwHWSHmq+MiJCUnR2x4g4HTgdYOzYsZ3exszMzGxh0qMWqIh4Mv//DHApMB54WtJIgPz/M3UdpJmZmVk76TaBkrS0pGUbfwPbAn8GrgD2yTfbB7i8roM0MzMzayc96cIbAVwqqXH7CyLiakl3AhdJ2h94HNi9vsM0MzMzax/dJlARMQXYsJPtzwJb13FQZmZmZu3MM5GbmZmZVeQEyszMzKwiJ1BmZmZmFTmBMjMzM6vICZSZmZlZRU6gzMzMzCpyAmVmZmZWkRMoMzMzs4qcQJmZmZlV5ATKzMzMrCInUGZmZmYVOYEyMzMzq8gJlJmZmVlFTqDMzMzMKnICZWZmZlaREygzMzOzipxAmZmZmVXkBMrMzMysIidQZmZmZhU5gTIzMzOryAmUmZmZWUVOoMzMzMwqcgJlZmZmVpETKDMzM7OKnECZmZmZVeQEyszMzKwiJ1BmZmZmFTmBMjMzM6vICZSZmZlZRU6gzMzMzCpyAmVmZmZWkRMoMzMzs4qcQJmZmZlV5ATKzMzMrCInUGZmZmYV9TiBkjRQ0j2SrsyXV5d0u6RHJV0oafH6DtPMzMysfVRpgfoM8GDT5e8CJ0bE24Hngf1LHpiZmZlZu+pRAiVpZWAH4Ix8WcBWwCX5JucAO9dxgGZmZmbtpqctUCcBnwfeyJdXAF6IiFn58jTgbZ3dUdJBkiZJmjR9+vQ+HayZmZlZO+g2gZL0IeCZiLirNwEi4vSIGBsRY4cPH96bXZiZmZm1lUE9uM3mwI6SPggMBoYAJwPLSRqUW6FWBp6s7zDNzMzM2ke3LVAR8cWIWDkiRgF7AjdGxN7ATcBu+Wb7AJfXdpRmZmZmbaQv80B9AThK0qOkmqgzyxySmZmZWXvrSRfeHBExEZiY/54CjC9/SGZmZmbtzTORm5mZmVXkBMrMzMysIidQZmZmZhU5gTIzMzOryAmUmZmZWUVOoMzMzMwqcgJlZmZmVpETKDMzM7OKnECZmZmZVeQEyszMzKwiJ1BmZmZmFTmBMjMzM6vICZSZmZlZRU6gzMzMzCpyAmVmZmZWkRMoMzMzs4qcQJmZmZlV5ATKzMzMrCInUGZmZmYVOYEyMzMzq8gJlJmZmVlFTqDMzMzMKnICZWZmZlaREygzMzOzipxAmZmZmVXkBMrMzMysIidQZmZmZhU5gTIzMzOryAmUmZmZWUVOoMzMzMwqcgJlZmZmVpETKDMzM7OKnECZmZmZVeQEyszMzKyibhMoSYMl3SFpsqQHJH0tb19d0u2SHpV0oaTF6z9cMzMzs/7Xkxao14CtImJDYAywvaRNge8CJ0bE24Hngf3rO0wzMzOz9tFtAhXJv/PFxfK/ALYCLsnbzwF2ruUIzczMzNpMj2qgJA2UdC/wDHAd8BjwQkTMyjeZBrytnkM0MzMzay89SqAiYnZEjAFWBsYD6/Q0gKSDJE2SNGn69Om9PEwzMzOz9lFpFF5EvADcBLwbWE7SoHzVysCTXdzn9IgYGxFjhw8f3qeDNTMzM2sHPRmFN1zScvnvJYFtgAdJidRu+Wb7AJfXdZBmZmZm7WRQ9zdhJHCOpIGkhOuiiLhS0l+AX0r6JnAPcGaNx2lmZmbWNrpNoCLiPmCjTrZPIdVDmZmZmb2peCZyMzMzs4qcQJmZmZlV5ATKzMzMrCInUGZmZmYVOYEyMzMzq8gJlJmZmVlFTqDMzMzMKnICZWZmZlaREygzMzOzipxAmZmZmVXkBMrMzMysIidQZmZmZhU5gTIzMzOryAmUmZmZWUVOoMzMzMwqcgJlZmZmVpETKDMzM7OKnECZmZmZVeQEyszMzKwiJ1BmZmZmFTmBMjMzM6vICZSZmZlZRU6gzMzMzCpyAmVmZmZWkRMoMzMzs4qcQJmZmZlV5ATKzMzMrCInUGZmZmYVOYEyMzMzq8gJlJmZmVlFTqDMzMzMKnICZWZmZlaREygzMzOzipxAmZmZmVXkBMrMzMysom4TKEmrSLpJ0l8kPSDpM3n7MEnXSXok/798/YdrZmZm1v960gI1C/hcRKwLbAp8StK6wDHADRGxFnBDvmxmZma2yOs2gYqIpyLi7vz3DOBB4G3ATsA5+WbnADvXdZBmZmZm7aRSDZSkUcBGwO3AiIh4Kl/1T2BE0SMzMzMza1M9TqAkLQP8CvhsRLzUfF1EBBBd3O8gSZMkTZo+fXqfDtbMzMysHfQogZK0GCl5Oj8ifp03Py1pZL5+JPBMZ/eNiNMjYmxEjB0+fHiJYzYzMzPrVz0ZhSfgTODBiPi/pquuAPbJf+8DXF7+8MzMzMzaz6Ae3GZz4BPA/ZLuzdu+BBwHXCRpf+BxYPd6DtHMzMysvXSbQEXErYC6uHrrsodjZmZm1v48E7mZmZlZRU6gzMzMzCpyAmVmZmZWkRMoMzMzs4qcQJmZmZlV5ATKzMzMrCInUGZmZmYVOYEyMzMzq8gJlJmZmVlFTqDMzMzMKnICZWZmZlaREygzMzOzipxAmZmZmVXkBMrMzMysIidQZmZmZhU5gTIzMzOryAmUmZmZWUVOoMzMzMwqcgJlZmZmVpETKDMzM7OKnECZmZmZVeQEyszMzKwiJ1BmZmZmFTmBMjMzM6vICZSZmZlZRU6gzMzMzCpyAmVmZmZWkRMoMzMzs4qcQJmZmZlV5ATKzMzMrCInUGZmZmYVOYEyMzMzq2hQfx9Aq4065qrK95l63A41HImZmZktrNwCZWZmZlaREygzMzOzirpNoCSdJekZSX9u2jZM0nWSHsn/L1/vYZqZmZm1j560QJ0NbN9h2zHADRGxFnBDvmxmZmb2ptBtAhURNwPPddi8E3BO/vscYOfCx2VmZmbWtnpbAzUiIp7Kf/8TGFHoeMzMzMzaXp+LyCMigOjqekkHSZokadL06dP7Gs7MzMys3/U2gXpa0kiA/P8zXd0wIk6PiLERMXb48OG9DGdmZmbWPnqbQF0B7JP/3ge4vMzhmJmZmbW/nkxj8Avgj8BoSdMk7Q8cB2wj6RHgA/mymZmZ2ZtCt0u5RMReXVy1deFjMTMzM1soeCZyMzMzs4qcQJmZmZlV5ATKzMzMrCInUGZmZmYVOYEyMzMzq8gJlJmZmVlFTqDMzMzMKnICZWZmZlaREygzMzOzipxAmZmZmVXkBMrMzMysIidQZmZmZhU5gTIzMzOryAmUmZmZWUVOoMzMzMwqcgJlZmZmVpETKDMzM7OKnECZmZmZVeQEyszMzKwiJ1BmZmZmFTmBMjMzM6toUH8fwKJs1DFXVb7P1ON2qOFIzMzMrCS3QJmZmZlV5ATKzMzMrCJ34S0C3FVoZmbWWm6BMjMzM6vICZSZmZlZRU6gzMzMzCpyAmVmZmZWkRMoMzMzs4o8Cs8qadWIv97E6W0sMzOzqpxA2Zuek0IzM6vKCZTZIqiVc4O1cwLq5NPM6uIEysysg3ZOCnsby8zKcgJlZvYmsCgmhW5p7X2cVsdaFPVpFJ6k7SX9VdKjko4pdVBmZmZm7azXLVCSBgKnAtsA04A7JV0REX8pdXBmZma2cFtUu6r70oU3Hng0IqYASPolsBPgBMrMzMxarpXdkn3pwnsb8ETT5Wl5m5mZmdkiTRHRuztKuwHbR8QB+fIngHdFxOEdbncQcFC+OBr4a8VQKwL/6tVBVteqWH5MjtVfcVoZy4/JsforTitj+TEt2rFWi4jhnV3Rly68J4FVmi6vnLfNIyJOB07vbRBJkyJibG/v346x/Jgcq7/itDKWH5Nj9VecVsbyY3rzxupLF96dwFqSVpe0OLAncEWZwzIzMzNrX71ugYqIWZIOB64BBgJnRcQDxY7MzMzMrE31aSLNiPgt8NtCx9KVXnf/tXEsPybH6q84rYzlx+RY/RWnlbH8mN6ksXpdRG5mZmb2ZtWnmcjNzMzM3oycQFnbk6ROti3RH8diZmYGbZhASVqy8YMpaU1JH5S00C96LOnbkpZrury8pG8WjjFQ0s9K7rObWNe1IhZwZofYy1B/7V2tJH1T0hBJgyRdJelJSXv093GZlSZpk062fag/jsW6JmmApI0k7SBpK0lv6e9jandtl0ABtwBLShoJ3AgcCJxVVzBJH5W0bP77WEm/lrRxDaH+KyJeaFyIiOeBD5YMEBGzgbVbkXDmWAMlDak7FjBN0o8gJZ7AtcDP6w4qaaikdWva/Yci4iVgB9LEbhsCX6wpVstPTCQtVde+m2Ic3nj/STpN0h2Stq47bp0kbdp47iTtJel7klbp7n69jDUqT0GDpC0kHVbT5/mnktZrirsX8D81xJmHkqUX9jh1v075++B04FHgOGAv4DDgekl/krSvpKK5Qqs/u5J2lHRC/vfhUvttxwRqQES8DHwE+HFE7AJsUGO8/4mIGZK2AD5Aau34cQ1xBjZ3O0laEqijG+ph4PeSjs4ftMMkHVZDHIAXgcn5A/B/jX+lg0TEV4B/S/oJKXn6fkTU0tIm6YbcMrQ8cC9wnqTjawjVSF4+CFwUEf8C5uuqLKglJyaSNpP0F+ChfHnDRvJbg4Mi4iVJ2wIjSI/pe6WDSNpV0iOSXpT0kqQZkl4qHSc7HXhF0gbAF0iTE59XU6zLgJC0JvAzYC3gghri7AacK2kdSQeSfpy3rSEOks7Nn9+lgPuBRyUdtbDGyep+nb5JOiFdMyK2i4iPR8RuEbEBsCMwFPhEwXjQos8ugKTvAJ8hrdP7F+DTkr5dYt9tmUBJGgfsDVyZtw2sMd7s/P8OwOkRcRWweA1xzgdukLS/pP2B64BzaojzT+BmYBhppvjGvzpcSfrw3QE80PSviPzDtaukXYHbgU2Be0hfJruWitPBsNwytCvw84jYBNiuhjjXSJoMbAZcJ2kF4LUa4jS06sTkRNLz9SxAREwG3ltDHIDGEOIPAuflWHV8p30P2DEihkbEkIhYNiLqanmdFWlo9E7ADyPiZKCuWG9ExEzSe/2UiDiSGtYzzQvO7wn8mvT+2zYiXiwdJ9sgf353Jn3HrgZMWIjjQM2vU0TsFRE3RydD8iPimYg4KSJK/1a16rML6bd9m4g4KyLOArYHinQht2Nt0ZHA14ArI+LPktYgnT3X5UlJpwHbAN/NrUTFX8iI+K6k+4BGM+U3IuKaGuLU1g3USawzu79Vn3Rsar0HWCxvD9IXcmmDJA0HPgp8pYb9AxARn5O0EjA9ImZKep10pl6X5hOTA/O2Wk5MIuIJzVv3P7ur2/bRZEm/BdYGvqRUG1fHvCxPR8SDNey3M/+RdDTpjP99uetksZpizZL00Rxr57ytWCxJ9zPv6zGM9J67XRK5haO0xXLX9E6kE4XXJb2xEMeBml+nBkmfAs5vlJrkVvi9IqKOFuRWfXYblgOey38PLbXTdkyg/hkRc2qD8tlLXV1QALuTMtITIuKF3MVxdB2BIuJ3wO/q2HeDpGGkJPSdwOCm2EXrrXKsNYFvAet2iLV2if1HxL4l9lPRt4DfA7dGxB05gf9b6SA5Uf8Y6cz1COAtwJrA30vHyo6iNScmT0jajNRKuBip6byu5GNfYBPg0Yh4WdKKwP41xJkk6UJSV8qcVsKIqCOB3wP4OHBwRDwlaVWgeLd4th/pu/V7ETFF0urALwruvz8Kxc8gfYb+TCplWBX490IcB+p/nRoOjIhTGxci4vnc5VpHAtWqzy7Ad4B7JN1EKpN4L3BMiR233USakv6Y//wZ8IuImNGCmFsAa0XEz3LrwzIRUeRHU9KtEbGFpBnMm2ELiNJdATmrvwr4dP63D/BkRBRPCiXdQurCO4F0ZrQv6TEVLRCVdA7wmQ5nRt+PiP1KxmklSecDfwV2j4j1lIpQb42IjWqOu0RE1NZVmL8ITybVE4pUs/aZiHi2pnh7kmo3vqVUbP2WiLircIzO6u2irvefpJVJ30c3SRoMDIyI/9QUa3Fg1Yh4tI795xibAg80vstz8fA7IuL2umI2xR4ADIqI1xfmOC16ne4ndU1GvjwQuC8i3llTvNo/u02xRgLj8sU7IuKfRXYcEW33j9SicTzwGHAusFWNsf4X+A3wcL68EvCH/n4O+vB47s7/39f4HOQ3TB2x7sr/39+0bVINce7pybZCsd5OWt9xcr68AfDFGp+7e5q2Ta7xfTGeVOz693x5Q1I9Re3vyRof0w+B04AH8+VhwJ39fVx9fEz7AXcDj+XLawPX1xRrB1IS/7d8eQxwaQ1x7iGfrOfLAxrfU/W40ywAACAASURBVDXEGp7fE1fmy+sCExbWOC1+nY4HLiKVmWyd//5+TY+p9s8usOqC/pWI0Y5deETEXyR9AbiN9ES/W9JM0g/Z5YXD7QJsRPrSIiL+oTytQUmSzouIT3S3rYCZ+f+nlYaF/gNYsXCMhtfymddjkg4hjRgq/tyR6neWjzT1Q6Obsq737hnAl4BGU/b9pOby7xSO83ruxmuc7a0G1HmW/ANSl8plkIq7Jb2/1M4lncICahgi4tOlYjXZLCI2lnRPjvFcPlMvStLapJG5IyK1Fm5AKiovOo9b9mlSsns7QEQ8LGlEDXEAvg68C7gpx7pX0ttriKPIv2g5zhuqbwqNs0kDdr6QLz8CXJi3L4xxoHWv0xeAg4FD8+XrSN+HdWjFZ/cq0ndSc0FmkJLft1CgBrTtEiileXf2JQ2fnAjsEqkWZRXgVqB0AvV6RISkxg9ZXfN5zNMMmr9A5ptgroDjlCbs/Dyp73pI/rsORwJLk770v5Vj1dGt8X3gj5Iuzpc/muPVYemIuK1RBJ3fGzO7uU9vfBO4GnibpDNJZ3wHLvgufTIgIh6vsbh7UsF99dTMnMA3PrsrAHUU8v6UVBd5GkBE3CfpAtJrWNqrkQqSgTndKHWZGanus3lbHTUdUyR9mrnTwxwGTKkhDqRuoAtyIT6RBmjU8Z5oVRxo0esUEW+QXqM6pvHpqPbPbkSs33xZ0ihSkvgBoMg0Bm2XQJG+rM4AvhpN/f6RRvb8bw3xLsqj8JbLBXP75WMoQtIXSS0aSyrNHdP4FLxODatQR8Sl+c97gHeX3n+HWLcDSHq9hpa05jjnSpoEbJU37RoRf6kp3LO5SLPxwd6ZNDVEEZIGRMQbEXFVfkxbkN4TX45S/fKde0LSeFJx90BS4frDpXYe5Yc598SpwK+A4ZK+RhoQ8rUa4iyVT+Kat82qIQ7AHyR9HhicWwg/xdzpXEp7UNLupBbe1UknQn+qIc4hpBbQY0mfqxuAg2qIA2kU4zDmfn7HAXXM2dWqOFDz6yTpoojYXfOPmgQg6hkt2arPLpLWAr5MasX7PvDpSNNC9H3fTS2rb1qStiFN7CbgmogovkSJpO9EC6YYyKOrTgHeGhGbSFqfNAt6HRMMjidNPDo0IlaVtCFwQEQcUUOs2gr9O8R5Oymx3RSYDjwF7BkRUwvt/y7g0Ii4o8T+KsR9C+lH7AN50/XA4ZEm8CwZZzjpLK/jyMyturxT3+K9k7kF69dHxJ9riPE74HDg4tztsBuwf0T8Vw2xBpKSiznfR8BpuXWgdKylSVN1NCa1vAb4eqT5whZKksaSBjG8E5hMmi9pt4i4d2GMk2PV+jpJGhlpxOdqnV0fEY+XiNNJ3Fo/u0qz33+Z9Bp9jzQoreiUKm2XQKnmofH9Ren0dRdSi0MAt0TEZTXEuRH4KnByRGyU494fEest+J69ivUn0rDryyKPHpP059KxcsvjWGB0RKytNH/SxRGxecEYo5qTJElDSZ+PF7q+V6/ibEZKcO8Ajon6JhRsxNs0IupoVegq3rWkWpD/JrU87EOa6+oLC7xjtRi/BQ4rldT2IN4apKR6M+B50rQWHy8ZP9d8nlD6C76LWHtExIV1x2mKV3sNmaQhkSa2bIxYewfph/kvUXBkXKvi5P23+nX6bsfPaWfb+hijZZ9dSbOBJ0i1UPN9rkrUZbbjTORnk6YwEPBfpJEAtb2JlJdl6PDvCUmX5i/OUk4l/aDcT5o75BBJpy74Lr2ybETc3LiQizfr6m4Y0MnZSR0/ALuQauL+A6nQn/LF6tdLOqZR3BoRL5ZOnvJ+byMNp/0LcJekk1TjMjjAj5SW2lmu+5sWsUKkCVZnRsTvIw31L9369DPgWklfVpprqlYRMSUiPkAqPl0nIrao4QdgFdL7odhJwQJ8QtLVhb/fFuSnpHUeZ0KqISPNTF7SPUrD4omI1yNickTcWzqpaWEcaP3rtE0n20q3srbys7sfaZT9HcBdnfzrs3asgVoqIq6RdEJEPAYcm2tF6lp88iRgGmltIZE+2GuSRuWdBWxZKM5WpLlPGn3m51Bw2ZMmzylN6taI82Hg6RriQM11NU1aUei/EWm0y12SDo+IOme/HwqsT2rNeIB6Cp8bxpJqJu6Q9I2IqGtdtYZGbcFTknYgjQIdVjJARFycu9X+hzTJ5Xk0PYcRUTQRzWeyx5NGATfeg3dHRLFFxyPicKVFzH8o6UFSi03zY7q7YKwP5dq+q5SK4TvGeq7LO/dOK2rItgJOUlom69Cob76kVsVp2esk6VBSYf8aSqtlNCwL/KFEjIZWfnZ7Wpcp6ZTelp20YwLVqqHxDTtGxIZNl0+XdG9EfEHSlwrGeZQ0/0SjxWaVvK20w0lzZ42W9BjpR3r3GuJAGu76A9LjeppUV3PoAu/RO7UW+gNEmuTvSEmbkNYsnEb6YDcmPC1SSCnpANLZ+InAIXXUtjTL+z8pd639UWlh38bQ3ojya7p9M3d/fo7UVTmENFqztNdJLZJLkL4f6nweHyC11l+bu1Weg/ILP0fE3fk751ekk7hGfUVQuBUvIi6T9DfSupn7d4hVusXjX7k0o5F87kaqLSwmt4TvIum/SIX4dzLvD/OOC1Ocpv214nW6gLRCxneYd4buGTUk09Daz25P9LrVtx0TqI5D44dSz9D4hpeVRjhcki/vBrya/y5ZILYsaTRFo3h4HCkDvwKKfsAfAbZUGhaq0kXCHbweEaWb4ucTEScoFfq/BIwGvlJTof9WpMLQM0hdrnV8sD8AbBERC/wBkbRORDxUImA+Wz6GVFB5aqMVpQ4R0Rgx9iJQbJ6pZpK2Jy1vcgWwcQuKnmdFxOcl7QHcIumTFB5GrlTk/33Sj+JWkRZXrYXS/GPHkr7r9m56zeryKVIN2TqSniTXkJUOImk0qfbuFur7/LYyTktep1yH+SKwV477FlL98TKSlomIYstL9cNnt1ZtV0Tearl/+WTSkP8gDQ89ktTytUlE3FoozvsWdH1E/L5QnEdJZxTn5C7Q2uQWrjuAs+pIaFpJ0i+BlUnN8ve3wfEU6SKSdBswFTgqapwmQdIPFnR9iYLNpli3kFrvFtgFrqbJV/sY756mQRLrkT5fq0ZEsbqy3MrwHeCndSa4OdZfSa1c34iIV+qM1SHu0qS6yeLLc0k6jrSw75ERcXXp/bc6To7V0tcpl3v8H2k1jmdI63Q+GAWXcmn1Z7eHx9Tr79q2aoGStDdp8dHRedODwA8i4oKa4g0EdoqID3dxkyLJE6QESWlG4eb1eJ4ptf8m44G9gYuV5p06C7ikpkx/LWA74KDcNfQLakjcJO0KfJc0e6yop/vp+ojodtZdSfv0tG+9j0p1EX0lIq7vNpj0xYjoy2zrh5AGR1xEqnsq3sXVEBHv6eFNbwBK1Ckd0BT7z5LeQ/oRLWl8REzv7kaSfhURH+ljrF2iB/Oo9aU2pMN+lgA+AowCBmnuJLVf7+u+m8wCNoqIVxd0I0nb9PFkr1VxoMWvE2li2E1J34UbKc1DVrSlsB8+uz3R6++qtmmBkvRx0ozZnyMVcIv0BH4POD4izq8p7h0RMb6OfXeIszupEHUi6bG9Bzg6Ii5Z0P36GPP9wM+BZYBfAt+O+ub02JK0tMEQUqvUF6PQXEe5Ve3DEfFgif318ViKFg/3d5xS8XKX8UdJ01rMIo2cvSRqGMlY4ZjmtBz18v5bRcSNOYGfT0T8uvdH1+tj6tNjqhirVCvo1aQuortoGqUbEd/v6757cSyL3Oe34Os0KSLGSppMShLfkDS5Q41wS9TxPpe0VGcNCZImRMTZvdlnO7VAfYqUcTe3Xlzb1CVVSwJFKgT8IekLv3nm82KjXrIvA+MarU5KEw5ez9zaqyKUTu+2IS2Hsx6pj/58UsJ2DbBOwVjLkVq7PkkqVj8SuJS0RM2FwOqFQj3dDslTVlvLSj/r0+OKiGeBnwA/kbQyaTTrXyR9Ieof+dflYfXx/u8DbgQ6a6EOoOUJFPUstVK3lSNi+/4+iKxVn9+F8XviBUnLkArWz5f0DE2/iS1W7H2uNPfeGaSGhMaEzwdHxGEAvU2eoL0SqKGddf1ExJQ8qqcuY/L/zc3JxUe9kPr+m7vsnqWeebgeIS1E+tOIuLFp++O566GkO0nJ7e4dWrb+JKnPo+SazvwnSbqQtBDua43r+6MFgNb9gNU+oWIHRR6X0lD8vUhJ/O8oNN9Kf4iIxtJRX48Os94rLalhPXObpPXbobaQ1n1+F8ZEdyfSAKojSSfGQ5n3d3FhdSKp1KQxYGuypPeW2HE7JVALKpKrrVI/ImoZKdSJqyVdQ6oTgtTV8dsa4ozrqvguIg4uHGvtrgpeI6LEYo3NZ/4vM3cpA+i/FoAiZ5aSro2IbbvaFhHjOr9nbfr0uCR9HdiBVLf4S1IXbl0TuPZUqVaAXzF/PcYl1LMYeHda2bLR1/dEY221QcC+kqaQToCKTg1iZd4TkdeelTQE+E2JffZB0fd5pLV0mzcVOUFtpwTqHZI66zYTUOsyLkoT/r2TeZeOKZp5R8TRuUVli7zp9Ji78G/JOC0ZuZBj1XqWFRH71rn/zkgaGAteTqNPE8spLf8wGBghaVnmflEMIc2n1V8u7uP9jyUNT98w//t2/sKq7cdSaW6haRHxWq7B2wA4t6nuaus+7n8d0vfC0A51UENo+q4oJQ9qOTci9l7Azfq8rIakjwK/6a4QmjQ6uS8+1Mf712HqwhxHaY7EZSIvJ5P19XVq7Ptg0oK+r9I0Bx7l5wVrxBsIjKApD2maMqFPn90OnsjdeKE0+/lnSCd6fdZOReRrLuj6uobkS/oJsBRpzpozSHNu3BER+9cRz6pTmrX9M40fRknLA9+PtExI6VhTSC0OP+vJCJhe7P9I4CjSiMKnmZtAvUTqdj2pdMwc93ukUTavAFeTko0jI+Lnhfbf6UKkDXUMXpB0L2mm9VGk1tzLgXdGxAcL7X8nYGfSMkJXNF01A/hlpGV5ipJ0K2keqDqWBmnEuJQ0eWCjRfyabk4a+hpvU+CBxvQFuYXjHRFxew2xOiv4f5G0HmixUc+SBpNm726sbXor8OMeJKW9iXUBaZTrbFLZxBDSWqfHF47zCPDuqHfuwEasI0jLrDzN3Hm06jrRWpGUZDYWLr6W9HvybJ/33S4JVH+RdF9EbND0/zLA7yoMt7SadTYio67RSLlVaE9SEf4A0jQQv+xwxtfXGANJIzCPK7XPHsS8NyLGSNqF1DJwFHBzq0fYSPpjRLy70L7ujoiNJR0NvBoRp9Q0eufdEfHHkvtcQKxzSQvUXsG8g1pKL08zhLTG5J6kOtDLSavVF5mPrkOse0iTJjZmIh8ATKpjlJqkq0hz+t2UN21JqsNbnVTLVmRAg6SLSIl04wTkY8ByEfHREvvvEKvx2d2b1JV8DHBX6WQjj5bctbORaqXlwWHvKpHE9Ke26cKT9DydF941ugCKrqfVpFF79bKklUjF3SNrilW7nAB+mTTqDuD3wHci4t81xHo7aZTfWyNiQ6VV1neIvs0n1JkBappYTdIwanrv5rPknwI/VZr89ALgREmXkCa06/PyOxExW2lB0pYlUMx9vnYALo6IFzvUBLRKya6vmZL2AvZhbr1csQVKJX0+Ir4HfCzHmUcUnBy0yWP53wBqXMIqnxCcA5yTp6DYDfiBpGERsUrhcGru7s/D4+v67RlEat16GkBp7r1zgXeRRpeVGhG6XkSs23T5JknFW6yzxXLX087ADyNiZk2f3S+SCv5vZ97BOnW8z58gtQzWrs4ejLZJoIAV+ynulXk4/vGk+aeC1JVXTA9rG0o5k9T/flC+/AlSK0od6+GdAXyJlEQB3E/qEiidQH2ftI5bo07no6RlforLr9UOpBaoUTl2YxqI31KuHu8mSTtFxOWF9tedKyU9RDphODRPo1G8u6EHSjZ570vq2vhWRPwtj4wrOWVCo05iUsF9LlBEfA26nrOmtPxjsitpUMswCk+rkk2R9GnSYriQur6m1BAHYJVG8pQ9k7c9J2lmV3fqhbslbRoRfwKQ9C7qe5+cRvpOnwzcnLvL60g+TiNN23E/9a9PNwWYmFsMm5O1oi2t2QZNdZFExPOSirRSt20XXm5laC7q/kcLYi4BDI60NlDpfdde25Dj3BsRY7rbVijWnRExTvMudVFXrHWZO7XEjXXUJ+U4U0jN/2d2rHGR9INSZ2O5xXUo6cvjFepvaW18pl7MLWBLA8tGjcu7dHEMRScYlLQkaVmVv5baZ9O+B5Oeo+kdtg8nLbRaR73Lu0knQctExHxz1hSKsQyp+24vYCNSd+EvgYl1DAxRWlvtB6TPb5Bmmf5syZqkplg/Ig3GaJxsfQSYBhwNXBmFRl1LepC0Ykaj6HlV4K+kSWSL1vJIWj2aptFQan56e6R1T4upqyyii1j/29n2xglE4ViTgS079GD8PiLW7+u+26kFCpgzIu5E0rpkzwJvAx6m4ASQncTcjLzMQL5MRJxbOMwU0qSdtdY2AK9JGhcRdwJIGktThl/Ys/msv1HbsDNQyw9yTpjqaiJv9snosP6hpM0j4g+Fm7Jb2uIqaSnSmf+qpNbJlUg/AHUvJDvfoRTbUVq76wRgcWB1SWNIdS5FFuYm/ehfzfzTZWxBmlLj0EJxmp1ETXPWNJlKelw/IhWQl2yZmU9OlGpfdDz7FClp2jxfPhf4VU4MS05Z08qJQeeZRiMiQmntztLTaPxO0kGkKQyaW4WeKxynuaV1mXy5eIlJk+YeDJG6q4v0YLRdAkV6YJsD10Zaj2cb6ul+AkDSecCawL3MnRsiSB+8klpS20D6Avm55g4hn00NK59nh5POlteR9DjwFHlF74XYD5h/zp9TOtnWJ7kV6INA48dxYtS7OOnPSMW0m+XLT5LO0oslULn78/puzvI/USoe8FXS2o8TASLiXqXFwUvZJCIO6rgxIi6V9M2CcTruv5Y5a5rcEBF1fSf0q5woXUI9XZHzhOoi/t87294bavE0Gsz97v5i07ZapjFQWpT7PFK3MZL+RTp5XeAiw70REedKmsTcHoxdS/VgtGMCNSsipksaIEkRcZ2kE2qMNxZYt46m62atqm2IiEmkhOYt+XIdCxY3Yj0KbKU0U7yiH9c966vcdbIZMFzSUU1XDQEG1hCvcaLQWCj785K2iIhjS8fK1oyIPRrF0BHxsgpXouak8A1JQ7vqBo+IPxcMObOTYviStRtLLeC6OlYRgBrnrGmyVuH9tQ21ZuFxgKtIyYVIyczqpC68dxaMMZo0YnY55p1UeAZwYME4AEREK2fXPx04KiJuAlCax+2nzD3B6zNJQyLipdxl90/mfteSB0v0uWWtHROoF3Oz3q3AuUrr8SxolvK++jPwVlLrSW2aaxvoZD2egnEeIk32eEv+V1sCpTQp5M7Mv8p6iVnIm+OsDjzVqDnJdS8jImJqwTCLk16bQczbQvgSqcm3tA+TFuycDSDpLNIghroSqNfz89bobl2Terp2/w3cL+k65u2qrmMkzwOSPgYMlLQW8Gmg5NxMz0gaHx0WxZY0DpjexX366hDSnDVvI7USXktqVS5pqVxE22kCHeXXAW2l79GChcc71s8oLWFU9Ls8DzC5XC2aRiMn7IfS1CoOnFZTF+/SjeQJICIm5rrMki4gJaB3MW+LYbEJQtuuiFxpHp6XSWd4nyQV2p4bhSf3kvQb0pO4LGkelDuYt9+3VB1FI97tpB/iK5oKrv8cEesVjrMMKYt/D6mFY1Xgzogo3rWWR1C8yvyrrH+3cJxJwGaNAvycuP0haljuRNJqUcOkj53EuQ94X1Nh4/KkwsZalrfIXeHHAuuSfpQ3ByZExMTCcfbpbHtEnFMyTo61FGnKjm1JX4rXkKaaKFLcLWk8cBFwNnPX9BtL+l7aM2qYCLIVJM0gTcjYWQIVEVFkHdAOLbmdBSo+4krSHyJi8+5vWZ6k+0sUJney3+GkFqdRzDtrd9GJhCWdQZoGpPFZ/QQwOyIOKBknx7qUdMLYGDX7cVKX+S6lY9WpHVugvhgRXyL9IJ8JIOnbpOHyJdXZLdipFtQ2QDrrn04qwH+e1KpS18i/1UongF0YFE2jFyPi9ZxEFSPppIj4LPBDSfOdVZROqElnyndLuoH0Q7Yl8D+FY8yRu8LvBjbN8T5T+qQkxzmnzpFxHWK9TEqgvlzT/u/ISdSngAl58wOkCQBradlt0Y/lo6WSpG40WnJHA+OYO5v7h0knrHVoycLjHZLDAaQaybpGil9O6k24nnoXGR8X806se2MewVaH/UjLxjRel1vytuLywK1fAJeXLp9pxwRqe+ZPlnboZFtfPUnqBppnbTNJW1BPd14rahsAXgAeIhVDHxH1Tv/wJ0nrlirIW4DpknaMiCsAlJbYKP3j3zgTakliHRE/l3QTaYK/AL4SEU+WjiNpnYh4KHcxwNz39qqSVi3dXdOCkXHNsdYG/pv5k42SycGzpCHjHym4zwVp1Y9l7ZrqPm8mzUTeWMrlq6QaojoMoTULjzd3888iPZ5fFY7RsFRE9HkNxB6YLWnNyMum5QEZtbwHc8t7Hd36nfk+aZ6z4yTdSZqy48oSLdVt04WntJDhIaSJCpvPXJclTVtfdBispCtJrV33d9i+PvDtiPhw5/fsdbzm9XgGkLobiqzH0yHOHqRh1huR6nduIS3Z0adFcLuIdT/p9XqUeVdZLzpiLdfrnE8aei/SLLafjAKzgneI08oJT5G0I01raUVE8RXQJZ0eEQflZK2jYt01TfHuIo12mVhnV3Xe72TgJ8zfhXxXl3fqXZyWzOGWY9Uyj1qHGNtGxLV1xugQ76+kyQxfy5eXAO6LiNGtOobS1GFuprxtzvQxhWN9E7gtIn5bet8d4mxNGq07hfQ9uxqwX0TcWDDGSRHx2aYSmnnUcaLVFHsg6bvpQGD7EgML2imBWh5YgTSL9TFNV82oo7lceRLILq6rpS+7lSSNIi2C+jlgZEQU7fLKMTpdADrqW/i59jlDWvVjKekUUj3SL/Om3YEHayq2bhlJf4qITTXv5Kr31VHbJemuiCg9F05ncVqyPl2O1ZIfyxzrfub/EXuRNKP2N0ud3En6Mun9fWnetDNwUcnBJsrL7uTPVWc/zEU/V/lEYcdGq7HSsk8/rKkGagawNKkU43VqGlmYE1tIXa6QGzIaiW+hGJtExF35+ZpP1LAWY467JKnreA9Sd+uVEXFEX/fbNl14uUnveeCjkt7J3LXc6hpJttwCrluydLDcHHoyqQYlgD8CR0ZE0SUNJJ1Pqjd4irT200GUHZnUbCXmDtt9ICJuKblzSR/PXV1HddgO1Dbtf6smPP0ATdNn5FF4JYf4zyMXrf+C9MNVS4Kb1T0yrtlvJB1G+mGuc+K/Vs3hBqlr/0uSXgNmUt8wfIDfkVruGsO79yRN3fBPUuF8kVb4iPiWpN8x9zt934i4p8S+m7R62Z1DgMtyl/XGpBP/D9YRKCLqfs81/DH3HtzX2JDrJov1KDS1Do+JiJObr5P0GdLarUUpLfw8njR57A9Jg3WKTHfSNglUg6RPkYo2L8ubLpJ0akT8qHCoSZIOjIifdoh/AHNH3JR0AWnNuMYogz1JP2jvKrHzpubjH5O+oGprQZE0ktTfH8x9rj6hlNl8JCJK1ZA1hrW26gsEWvdj+TfSbPtP5Msjc9y6NM6+LpL0BnAhKZkqNvFfdgSpqPs10vv7GuAbhWM0NEb8Hd20rfjEf9HCWZNb+GMJ8IEO3e33Ky+1I6n0RJtLAS9FxM8kDe+sC6wvGt3fUcNozy7i3am0vt+1pJHIH4gOS/6Ukr9X9wZWj4hvSFqF1KtQpBBf0ltJ02YsqXmntxjCgudC64t9SA0KzSZ0sq1PJA0gTZK9V+QpY4ruv1268BrymfJmjS+p/KV1W+kuAKVVui8lNYk2D1FeHNglCq8R1lk3hqTJHUY99GX/RdcY6ybWr4HfRsQZHbbvR2rW3rkVx7Ewk3Qj6azoT3nTu0gjk14EiIhdu7hridhrkUb87R0RxScJXdSow6zJpAEMtcyaLOmGiNi6u22FYk0GDmz8ECvNb3VGRGyoguuiKa17NhYYHRFrS1oJuDgKTjfQVU1NQ6namk7irEtq7X++ZJwOMX9MmiB2q4h4Ry53ubarEpRe7H8fUvIylnlb8GYAZ0fBEYxKE/l+jFT72dxjsSzwRk3v89rW+Gu7FihS9tvcetJoxi4q0ordm0l6P9AocL2qZMEcgNIsqJDWGTqGVPMSpNaA2uscavLOzn7gI+Ks/BiLUovmQcmxbqLzGorSw76LrMVUhdIq7nvkf7OBzxfcd0t+wHKsrSLiRs27vEVzrNIjrloxa/Jg0tn+ivkHsrkV4G2l4nRwAHBWPkkVadDJAUoTGn6nYJxdSINa7oa0MLzSfH8ltWpampZPf0OaNmNjSfdAKndRwWlccqvdOZI+EhF1jSRsuI2UcK5IGh3XMIOmrsPCbpD0EeDXjZKJUtomgZI0KCJmkc70bpfUeCF3Ye7EXnX4EHBWHWeTWWMW1MYX4sFN1wXzrjvUF2vklqFOFW7R6LTVIjc117HERSuHdv9309+DSQuTziodJCJuyInhONL7YFJdXQAAShO5LkZa/+6jpWvvmPvDsitpZv+f58t7AU8XjvU+4EY6r9GpY8h6K2ZNPhj4LKmu8C7mfl+8RKrbKC53+a+vtBQTMe/yOxcVDPV6RITy/Go1PHe1FR93FUddrI5QU9iZeQRZ4/kbTtklixrWy/XH84iIr5cKEGmS4scl7Q38o8PztzJpoevSDgaOIk3T8AoF6wrbpguvuQtKafK6LfJVt0QNQ0Ob4h4A7EtKJn8G/CK6WMernUl6hFTY2KmIuKFgrB+QujqPjIhX8ralSD+ib0TE4aVi5X3XPrS7m/h3RMT4wvvcF/g6qWhSpPf7V+qq5N+4dwAAIABJREFU4ZA0Omqe2DLHmRQRY7vbVmP84mfRatGsyflH8ksRUVfNWMd4S5BOEEYxb8tusR/MHOe/SevvbUNq2doPuCAiTikYo7MRhXPUUALSytUR9mbu6LFzSCtaHBsRFxeO87mmi4NJjQsP1tTS37Lnr07tlEDV1k/Zw/ijSYnUXqS15H7afNbZh/22pLuhxTVQi5Nm0t6bNGoNYHVSofwXouCw1xyvlUO7hzVdHABsAvwgCs9ZozQ3zhaNVqd8Vnlr6ThN8YYC/8vcda5+T5rgsujJgqQHgR0aLVz5TP23EfGOknEWEP/vEbFq4X0uT5o1ec5JHfDVyMvwFI7Vsu9BSVeTau46zqP1/S7v1PtY29C05E5EXFd4/6st6PoovDxTZyd1JWtaO4m3DrA16fm7IWpe6y/HXIL0Wm1Zw75b9vzVWYTfNl14wHAtYO2kqGfIOjDnzG+d/O9fwGTgKEkHR98n8GxVd8MT3d9kbkLXl0D5rOGzko5l7sruj0SH0UklYmWNod2vk2ri8mHUMrS7uct1Fmm03P41xHmONGt8wwt5W10a0yTsni9/gtTiWrpY/UhgoqTmyfgOXvBdiqqjXrKVsybXVq/RiZUjYvuaYwBpKSGgaNLUYf+1r1/ZQe2rI3Q4mXuGNKp1znVRfrqOjpYidavVoRWrSzT8iFyETxoR/G/SiPg+t3a1UwvUU6Qh+F2tEP61muKeSEpubgDObM5KJf21rhaB/tLilqqWxVrYSDqbNHjhMlLCtjMpwWkUiv6gcLzOzvhq6RrNZ67r5IsPlW6R7CZ2sRYo9cOsyZo7aeIs0vD42uaBknQ6cEp0WI2h4P5n0Hm3WvHHJOnWiNiik5h1TTpZ++oIkv7G3JO5VUkj/USaw/DvEbF6qVg5XnM36EBgOKmVungNXtPz97Yccxo1rC6RYzWm5mie3LdIa1c7tUA9VbrvvYfuI/Un/6eT6/pc9yLp7IiYkP/ep64alyqHtDDGUlr2pNH9NDEiriy17w5xPgWcHxEv5MvLk+YQKT0P2RP5X2P236vz/8MLx2l4RdIWEXErgKTNgVdqirUJc+tqNpRERJxbaucLqHcRZQt5W7o+IrR8HqgtgAn5h7p5KaYi9UKtfCwRsUUrY0aajHZT1Tg3WCNBkvRT4NJGCYOk/yKdcJX2oaa/ZwFPRxrYVVwrnr8mtRXht1MLVEtroDR3cdVORaFFVjtkvf3eIrMwtkBJOo7U3Hp+3rQXadRaqRGMzbE6a6np1/q8EpQW9T0HGEr6oXwOmBARRVdbl3QesCZp8rpGXU1EwaU0Wl3v0mo5aV+LVMgLQETcXEOcTp/HUs+fpCER8VKHrqjmOLV1QeVBLesCUyOilq6h/2/vzMPsKqt0/3sJSEAGB7zQgoACRhEBmQyKIpPdiFO3I6AiTtAqBLnXayMqU7cTot1iN4oiAioiFwcEjSCzgEQRDBChtUXEtlWgmRoEQnjvH+s71KlKVSVFvr33OVXr9zz1VJ19kr1WJefss/Ya3iVpL2ITQ///U/UkgMZZLTbesYr2/hejf6faYrsodBg/CjzV9p6SNgd2tH1SA7Yaa8IfpAxUdQGtZTBZo6SJemkNBiNCHW5eRkj/PwIg6RSi1FU9gCJWkKjXf1LuXJrYI3g+45eFXjrOH19hbF9LZIPWKo/vacIOIcb36IqaJuj/gC9BwGa2f6QYha5+TSvZuiOJfq6VGcnUVFU8L7beQfT8bUAEoXOJtU/VdMh6gQ2hvdMkXyeyGmOlXKCyYnzJUH+WuDH4ENHj8idgY0kfqJ35l/R5okdoF+BLxIdyFWXwcfhD6TftSYPsC/yhtpHyb3gcUZb8M/F6/yUj67pq8hWiB/Pw8vjfie0I1QMo219T7C7sxRivrtWEPzABVAsNcWPt7dKSqQ0UY//q+7nfj7aXxy5Xs/kA2noCI03Wa1c871jmA2dI+kJ5fAAj5bWafKjv557eVPVeoYkGM9TcPsHrCR2oWut8JkTSO4ldj08isl4bAJ+n/s3YSURz/KhptYaYR2Rbf2J7lzJ9VW3pbqGVwMb2y8v3qr06E3AMMeW3NnARsKXt35RsygXU1xJ8ge0tFRsmjpJ0HLFbsAn2JiZoe8uYLy3HanMMEbD/yPbzFCLTtVf69FjH9jclHQZg+2FJVd9bJQu52PZi2zcqdMheRiwGn14BVJdIegFLa6HU6tno39PV+KLLchd+CLCR7QMlbUrcof8AwParhtDWx4BrFCrhInqhqiueFz5ABE1/Xx6fT9xhVsX2VWMOXaIQu6zNp4hMxg8Y6XNpknWARZIWMHrBb/WGa2Jn5g7AVcXGr8oHZm3u7r2mW+AB2w9IQtKq5cJfdZCl5cAGaKUs+Yjtfy+2bnaR0bD9Z0lN9PH0+gfvV6ymuYPYZ1mdklyY18S5x7DY9h2SVpK0ku2LJP1zQ7buk/RkRvqS5lLWWFVkPjFB/avy2XQl0QbycsXu2BWuYMz4AGqing2gSgC1vKljScfbPqiCyS8D1zGiWfMHQn26iQ+Axm0pUiU/Ju6MemOnH3DlXYU9bD9SJuQudIPCk71SWqGnN/XEBkw9j7hb3YvIOJxO6Mg0VWI7sqHzjseDth/qZdMkrUwzJfOLJB1LSI70B4VV+iTH8HtJTyCmM8+XdCfQWE+XpC1Z+uaxqpJ7G2VJYKUSpK0EPKLR63Ca2I5wTvl/OpYQWTWx3qc6kp5JbEjYmNH/T7XXS91VmrovBb4m6c/AeMNVNTgUOBvYRNLlxPDM6yrbeKLtX5Wf9yNEsg9S6BheTYUWkIFpIu8KhfBfoz0by+lHrYbrn9nebkzzelPj6q3YarJhchxbryQuio+z/fTSfH107QyKpFtZWm/qKDe4kqJkWvcGdieC0LMbstPfl7Q6MMt29X4bSZ8k9LPeAhwEvBtYZPvwSf/i1O2MJ6jrBj7AxtrdmShJzXdRbK58/i8DWwI3MDKVZFdWnlZMTfbKklv3ypKuuF5K0m+J32G8DGsj/Wp9tlcFZruhDRaKpc+fZ2nB06sr23k8kVlbieizWpuYSL6jpp1ia1Xid5lD/J/dBKzkipInpby6Zfn5cuBY298pj6edjEFXtNaz0RIPKRaT9lKjT2f0cuZhtPXzknJtbKVPH0cQZaGLIZqvy+9VFdtPq33OyVCM7j4PeC6hufLnhuyM7Utan2b6kiDKuG8nsqAHEMu5myi3Nt4vWd5HBwKbEr/PSU0G04W5tjdv2Aa0U5bcuOb5JqPcINxn+/ZSetoJ+DWRNWyCh22f0NC56Q3NeETK5xHG9Iz1D9ZU4sqSMHh0B62knxOTcrVYKOlTRGVkU+C8YucJtQzM2ABKI+J4a9Jez0YbHE3UfjdQTKvtTDNK2m3aej6wr6RbiJRyVb2aMSy2fXevLFSoduGQtC3we9t/Ko/3JdTAbyEyXXdN9vcfg723Eerjs4H/B7zediPBU6GtviQcU5lfpKHSSQ+1M3J9CqGyfxmwJzGC33Tfy5WSNre9qGE7rZUlFROT19q+T9KbiA/kf3E9aYYPA28FLOkbRDb3YmAvSS+xfUgNO2P4nqR3E03k/Z9RtQavLpJ0FvBd90kWlFLXTkT56yJicm6FkLQecVO1mqTnMZIxXIuYaqzJO4n30IbAS23fX45vTiVttxlbwivp8Qlp4e5vFKqoNVSyDS8gXpxXNPmB2YYtNaxXM8bWScTUzj8Qk3EHA6vYnnBR8xTPfzXxZr5D0k5Ez9ghwNbAJrZfP+kJpm7vESLL2vu3GvWGb6A0eZXt5/dez6Uv6ec1g121vzj2B5SRa9tbld/pmppl5f4ydTn/ghol/WXY3JnoQ/kjDQhpTmKzybLkQmArojT5FSIj+Xrbk17vp3D+RcR7dXXgd8B6tu8v/2fX2t6ihp0xNm8e53C1smTJfr6Nsi+OKIuvRpTyzgP+zfY1lWztRwSg2zF6qOpe4Cu1+++KzXm2/2VZxx4LMzYD1QuQJH3C9gf6n5P0CWLZanUkrd4XCfezwv+ZfewIvJD4kFkCfK/iubuw9Y+239x/oDT/v3mCP78iHERokzxINFz/kBjvrcXKfT0FbwROtH0GIZ1QVdSy0JZcR49LJH2QuMPcg+hLqv2aePmy/0hVGh+5ZmTHY+/8lU8/LicR76HrqKTM3I/GF9DsrY1Zg2Z2Pz5s24rdap+zfZKkmlnxB0rg95Ck/+hdy8v/WSOtEm54WtL2A8S+uH+TtAoxSfuX2tnwYusU4BRJr7F9Vu3zT8B+LP35+tZxjk2ZGRtA9bEHMbrez57jHFshSgPvl4gLx4aStgIOsP1uANtfqWTneCJF+Y1y6GBJe7gBvakWbY0SclOIW25b2QYA5YJ4OHC4YpLnrsq1/1mSZtleQvQF9We2qk8LtZ1JZXRf0ruAc21X7UvyaCHN9YiSoYGfupnpzDZGrreS1BM3FRGA3sNIVqiJxdm3NTVIULid6LfryQg0JqTZx70l0H0T8GJJKwGrVDz/EyT9HfG7rFV+pjxuRJ+uDGIcCmxo+12SNgPmuIF1VrYX004/8AWSPs3Ieq5LiBaGau8rSXsD+wBPl9T/Ol+TSsH7TC7h/T1xd/wM4D/6nlqTKEXtW9neVYRa7dkemVi7vnbKV2OmCkuwcb3tZ9e004atciH8IJFO7mXtRDSqn+iKq1wkfQT4ZmlwXZWQYtiKyKrtY/tHFe3sDtxGNDY+zyGdsAlwmu0X1LAzjt3xyl53E2n0f1zRSZtyx7+B7X8tjxcQo8kG/q/t/7ci55/A5juAjwAXEq+LnYmL8Jcr29kGOJ5Y/nw98Xu91vbCmnbaRtK/EQK132N0b02VMopCQ2gX4HIim/vjyjcj49lcj/jQ/KntyyRtCLzElXT9JJ082fO2969hZ4zNM4gJvLfY3qIEVFe4gcnqtig9V9cz0qz+ZmAr153M3IgoSX6M0bqB9wILXWHP30wOoNYmdHeW+set2JzXb29Ub0g5VmWUcoyd7xOZrVvL4w2Az7uI5w2jLUkfqxksTWDjBmCLkv5/F3ER3g14JnCK7RVeLN1naydCdG++y3i/YrR7DduNiK0qRv6XECrUEOXD1Yn+l51sv2IFz3858Ma+18K1hM7PGsDJtqtP4Um6iVCEvqM8fjLxwVJlwkvS9sCttv9YelwOIPriFgEfaeI60SYTBAN2RRkDRS3yJYR8xg5ET80Jtsfr66lh7/FEmW2JQj/pWcAPSmZlKNH4cjHVPzvaROPvHG1Ebqece6mVT64grTJjS3glVXg3sHfJnKxL/HusIWkN11+geGsp47nUmedRSU5+DLOBX0r6CXH3PxdYIOlbADUj/BZtLZC0di+9q5joeYmLpkclHuq7O/5rQnRtCfH7VX2f2P4xxIeLYkJtZeCe8tUUu49pSr5ORXtMMa20ojyuFzwVflwCjP8uH2pNcAej97ndW47V4gtEthBiUOJwokdua+BEIqM8tDSRLRnHhokpr2uIoP0Y4Fc0Nzl5KfCiUn4/D/gpsUi2dkWhtWW4RL/VaoyUkDehgbVPLfMXSTv1XQtfyIi6e1XU4MqnGRtA9ZD0XkI9+U/0ickRUxw1OZBoWlsf+E/izf2eyjYA/qmBc3Zt6wjbvT1Q2L5L0hHU1V15UNIWxOtgF0L5t0ft8dpeCfkY4gO//3XXlC7PLEk72F5Q7G8PzCrP1Vh1MUpF3fZ7+x4+pcL5x+PXwFWSvkv8272K0H45tPiwonv+ZvVlmd5AlI3PAs4qGbahRjF99Xaix7B/xUqVDFQJnF9F/Ns9hVBy37aBm9NRZh1TcW8npsc+qWaGM75CS8twCW26+cDTJH2NGNp5a62TS7qXyadam+i/OxA4tVSCAO4kmr2boDFplRkfQBEj5HNWtAdkWdi+ncp3QRPwTODrNZvxBsDWeM3VtV+78widpKcAn+mVGCS9DKgywjuG/w082/ZtDZx7PN4BfFmxqkFEtusd5UPuYxXOf5Wkd9oelVmQdADNban/D0b3L363fF+z0vlnSVq59ErsRtzF9pgO187TgBuJjOvRxPWpZlb8z0S26Rvlu4HtJG0H9VfGFCRpR+J36U3fNbHKpY3JTMq5z1eITM4l3rvzyudJrfOvCSDpGKKB/LRiZ18a2O9XGvvnOCRB1io+NJl9b2zl03S4CKwot1J/omYpFEKT81xGQ0uK+bia/QaFjQjl7quAL9dqfu7Y1s8UExv/Wh6/h2iqrIZjue+zxjn+fULhuja/p5kx7nFxqLg/t3fHNybo/WYFE+8DviNpH2I3GMSk5KrAqyucfylsH9XEefs4nZBluJ0oL1wGoFhM2sYNStNsavt1kl5l+xRJX6f8jpU4k/igmlO++jGRkarNIcSOs2/bvkHSMwgRyNq0MZnZz86EqKWJqcJvT/7HHxOvHNNXdULJ3n2kppEyNPN/iaGdJgOnHpeoIWmVGdtE3kMhnDgHOJfRkygrmv4fa2cpoczxjlWytRIhxbA/MUl2OhHg/HYYbZUsyYeJfhQD5wP/5JHVA9VourdBUk/iYUtiO/05jH7dfbaGnXHsrko0QG/M6IWkR1e2sysjshM32L6w5vmLjX+2fYhGtgmMwhXFQcsH418B5/Veb6U5eQ03s0y4NSQtsL2DpEuJD5U/EgKeje2Nm8CP/bycS9encM6J9PZqnb+1yUzFtOSmxLUVoiT6H7artoBIuoK4Sf0G8b7aG3iPG5gMlvRxQubiDPoWFjc0wLUSkY18KZFZ+yHwJVcIfjKAil6apah9d1si+ZfYvrM8fhJwiRtakivpOURQ8wqisXIH4PtNTLO1ZUvS45sImsbYaFR1uqTJJ8T2h2vYGcfufOIOeexC0uOasNckkra1fbUm2Cbg9rWvhhKFDMRZRDB/MjEx+RHbn2/ZjyqL1Mu5diT6kNawvZTeXk3KteHRZbhNTfpJupEo9/eyXSsRNydVpWkkbUz06faEkS8HDmnoxrtRdfW2mPEBVI/SG4Lt/2no/G8hNI3OJN5wryWyKKdVOv/KpQ7/HqIZ7x7iQvIt2w+WN92va7xA27RV7D0qQtrCRfGntrfX6JHhxsZr20INaI51jRpc0ZC0R81MvFrS2yvnfQFLZ3Sr6E2NsXMOkQm6pTzeiFBZXyHpkTE2ZgEH2/5MrXMOCmXC70ii5WRleFScdoU/n5porhsqJG2hGLG9AbhB0tUlo1KV8sb6O2LK64/A39UKngq9Rt2nAnvb3t326bYfLPYfAWqVNtq0BfAZotH1jnL+XzCiYFubVnobJM1X31ZwSU+UdG5tO31cIamRbGeHjDe189a2nRhWJK0r6aSSdUXS5qq79mR5qXoX79FyGtCXca2FYpXUp4i+pO3L13a17RTWJORULpZ0EaFDtpakszVaYfsx45Bs2bvGuSZD0maSvivpekmnS1q/aZvEzf2nGfm/2q58X2GyiTz0XA61fRGApJcQGiVV6r6S1rJ9TynZ/ZERIUMkPalizVcAtg+f6A/Yvn4IbfXOd6tG7wdrZOKFWJlwNrCJQhzyKTSj97Oe+3ZN2b5T0lMbsNNjJ+CtJXXeyuLYplALKxpmCF+hvVH8yai5+K8tvb3t6NvC0DBVm7gn4XJJn2PpvqSavX5fBk4lWj1eSfSR1dQmHI+7bf+giRNnAAWP7wVPALYvVl3hv68Ty0+vZvSdlqi7D+opKvo341G5Kb5NW9DeRRHbPy+9NU33NiyRtIHt3wMoVk40yZ4Nn79NriDGrdcB+nu47gWGer1Ky7Q2ir8MLq94rrb09q4H1qOFvXG2L1FDStpj6LUp9A+WmNgoUIs1PSJ1cqxCnqFpLpJ0LDH12T+ws8K2M4CC30j6MKF9AbGE8je1Tu6y1sQNb9QmRBF7Gj9N06YtaO+i2GMHRnobtpHURG/DR4g7vt4et5cAf1/ZxqMZUEYrdg81pRfkFmDHrn0ZchotV092kwUjN1oeLbq6Qrg9vb11gEWKnY/9H8o1WxeAZpW0+7G9S83zTcBsSc9j5LNjtf7HDU22Pr987y+xVgkMZ3wTuUKP6SiixAGhg3Jkb1quop2ziTHU7zYxXltzkmWQbLVN6W3YBLiWkTKhbR888d96zLbWZSQIuML2nxuwcY7tl5fSnRkd9A7d1AtMqpzcK0s2oZw87Wh6FL9vwnkO0XPSK7e+gpBLqLFCqGfreCZX0676/m1zAlSher8DcFVfY/x1TUxwS9qLpZXpq0mdlB6uibDtmtmu3sTia23X0LpbihmfgSqBUvUPx3E4jtDv+LiknxJaG+fYfqDS+dvKBrVmq+2LYqHN3oZdgE1s/5Okp/XG82saaDED2houysnJY6d8sMwmBBobKVe7SMEodKa28cji7CMJ3b2aNLKEeyJalspoTEm7H0mfJ9ZW7UJMPb+WylsEWspy9dt7VLSzifPP2AzUsqYXmkjFFruziNThO4G/qXW3XLkhfSBsSZp0N5Iri+8Vm2cS47yN9jaUZs1VgBfbfnYZMvih7SrTIRPY3JKlx66bUINuhYn6xtzsrrVpQ035gGXYuQnYsjelqxB1XWh7rDr5itqZBXzC9v9Z5h9+7DZ+bHuncbKgjWU/JX0SuAt4C7HM+t3AosmGeB6jnYW2t+z7vgbwA9svqmmn2FqdGNjZ0Pa7JG1GrHc5pwFbjYl2zuQM1I7EGpfTiSWDjWdVSvPfK4hM1DZAtQCgreCpTVuTBUjlLqwJ2upteIHtbYqEBrb/W9LjKtt4FElfJgQTb2D08uKhDaAYncWYDTwduIkRJfRkci6Q9BpCv63JO+lTgQWSvk1cZ19FTABWxfaSovnTGLZ3Kt/bzIL+A6GkfR1wALFa6ksN2PlL+X5/mQi+gwZ24RVOJgaretPu/0loJFYPoIjPWxjdN1tlgGsmB1DrAXsQ2hf7EBfj023f0IQxSd8k6tjzgc8RKuSPTP63Zja9u73y82m239z39AIiCK3NkQ2cczwWlzJKr4H3yYwENk0w1/bmDZ6/dcb2gJSenuriqtOYA4gswMOSHqChLEopUf8AeBHxet/fdhMLugGuLdWFMxmdbahyo1AyxRPSxM1lKUN9B/iOm10+fo5Cm+5YYp+laSZQg2hdeEORJMH2/RqjU1OLJtsXZmwA5RAOmw/MLynlvYGLJR1l+3M1bZUPymsJ0ckuxoSHlX45ibFZhabebI32NqiouBM7p84iJCGOAl5PDDM0xZWSNre9qEEbnVIkKJ6/7D+ZQOtZlCXEDYJp9kZhNpE56W9Grplp7cnRjHf9qSlLQwkojgDeSxG9LjITx9ds7O5hu7dm6iyF+vlsj146XpOHSkWmdwO5CX0Z/5o0WS6csQEUPFqL34sInjYGPksDW67LHcTrbH+09rmnOZOVFaqWHFrsbVhANNSeKulqYkGygNe5svjoGE4lgqg/MuRCmj3GjMmvRGQk/9CRO0OHpAts77asYxXszCN6Ps8iXndflXSi7eNr2gGwvX/tc445f5vDGO8j9tJtb/tmAEnPAE6Q9D5XXruiWM31Ndt3OVZyrS7p3bb/raadwhFEAuNpkr5G/J5vbcAONFgunMlN5KcS47vfB77R8IcXkj4FXEnz/QbTBkm/Af438eF4LNBrDhXwSdubdOXbY6Wtxt1x7P6auAu7jr4MgMt+rWFEoxeBPwz8Fjir4mTrtETSbGLa6iJCf6yXTVkLmG/7WZXtLQR2dFkErhAqvrKJ4F3SM4ETgHVtb1EGJ15p+x8bsLU+I/vVALB9acXzXwPs4dC26j/+FOC82tcRjbPzs8nrVWlbmEu8/n4y9vesaOdntrfT6P2mv7C91YqeeyZnoN5E1MjnAQf3lV+bmqbo9RsskfSXBu1MJy5hZKfeJUQDfo9qFypotbehbRX3HrfZrrI3a1DojcknU+YA4BBil2W/bMa9RH9mbcTo1UtLaG5o54vA+4EvANheKOnrQNUAStIniObkRfTpxVH3urTKeEGF7dsUGxlqM0uSejf4ZaqxkcEWSX8LXGj73PL4CZJebfs7DZhrrFw4YwMo260uUm6532BasLzpeEn7VZA0aKu3oW0V9x7XlA+S7zF6unDopvC6kiCZRlxB6OK81vbxRS7kNUQG7+uT/cXHyMnAVWUKD+DVNLdvb3XbC8b0Iz/cgJ1XE300jfTtFB56jM89VuYDZ0j6Qnl8QDnWBEfYfrRdxvZdJaPcRAB1JEuXC6uUemdsCa9tSkPgvsDTbR8j6WnAX9muKlQ2E9EQKaN35aukk8c5bNtva9uXFUXSbUwiQdL0IMCwo9g/tnuRzngxIep7ELEL7dm2qy/PlrQt8cEFcFlTU3hl2u+9wJlFJuS1wNttV90FWey8zvb/1DzvGBtL6Jsk7H+KaPCumoUqw07vIvoyAc4HvtTE4FNPa2rMsUbU1cu5GykXZgDVEpJOIHpPdnUIJz6RqGM3Jpw4U6hdp2+yt6GrHqjpRCkt9CRItqRhCZLpRn//h6R/Jcq7R5bHS/XBVLI5C1iX0e+p6oKnpcn6RKJh+E7gZmDf2r1+ks4CtgIuYHRGt42tFkNP0aW7i5hGhtBoepLttzZgq7FhiRlbwuuA53u0cOKdalA4cYZR7S6ghd6GqhNOy0tpHH47S++5GroMVJsSJNOUWX1yGrsRWYce1T8TJB1ETF39iZH+JxPBb1Vs/wbYvTSqr+SyPqYBfgRcTPweDzMiQpksHwcBHybUwSGyXVUXxPcNS6xTEhb9wxLr17CRAVR7LC53Yb1GtqfQrB7KTKJmP1GjvQ0Vm9GnymnAjcBfA0cT5eRfduTLCtOWBMk05XTgEkm3Ex/8lwFI2hRoQvdnHvGeuqOBc4+ilGqOIJbDW9KPgaNr2VZsQPgo8DbgFuLasyHR5/XBGjZmAmUi8x8aNjN2WKL3OXEPlYYlsoTXEpL2ZfQKl9cCH7J9ZqeOTQMkfc72eyudq/Hehi7olQ41sucXhs7uAAAXGUlEQVRqFaIXZW7Xvk2VtiVIpiOS5hJrOs7rkxd4JrCG7Z9XtnURMY7fRDP3WFvnE9nir5ZD+wIvsb37xH9rSuf/DLAm8D6PLEdeC/gUcL/tQ2rY6YKiVXjmso6toI1/tn2IpO8xTuWgiQEQSQc1oTkGGUC1iqRnMVLCudD20GYA2mCycX9oZuR/uvY2SFpgewdJlxLrTv4ILLBdTTm5LSQ9wkhzbSsLXZOp0/f+fQ4wh+hV639PNfH+vd72FmOOVWtOlvQr4JljtfxKdeFG25vVsNMF4w241B56kbSt7asl7Tze800NgEh6AUsvUj91Rc+bJbyGUcjIL7a92PaNkgy8DHg2Q1xCaYme9MMcYHugN77+CkLRuwmma2/DiaUP4MPEv+MawEe6demx0bYESfKY6b1/f1e+HseIrlBTd+7nSXojIdMAken/YcXze2zwVA4uKdf2oUPSnsRn0vqSPtv31FpUloCw3dMdezJwbsMyEABIOg3YhFin1t/XusIBVGagGqbc8b/d9q9Kj8EC4GvA5kQG4LBOHRwCyr/hXn0p8zWJN9+LK9qYtLfB9uJatpJkJtFGaajvvPcSOzR7/aUr0ZetXNHspGKp77fGZi8kvQl4/TBqkEnaipCwOJrRN1X3AhfZvrMBmycT+wovJRrJ5zdV4pX0S2Dz8QLfFT53BlDN0p8+lnQMMar5njKBd3VTuhfTCUk3AVv27lZKA/FC23Mq2pi2vQ0AktYlAsSn2t5T0ubEeo2mBA2TBGinNNQWReLkW0RmupdN2Q5YDfhb2//ZlW8rSt9kZlv2VgH2JHqDdwLOt/2OBuycCRxs+79qnztLeM3TH6HuSux0w/ZDpZcjWTanAgsUSsYCXgV8pbKNlzOmt8H2PZL+npheG+oAivj3Ohk4vDz+d+LOLwOopBHaLA2NsftEYDNGy3VUkSEpAdLzJe1K9HYBfN/2BTXO3wWSvmn79cS2gvHKk40sHLe9uAztmAhAXw1UD6CAdYBFkhYwugdvhbOFGUA1z0LFIuE/AJsC50Hs/unUqyHC9j+VN9qLiDfb/q6vZDztehvGsI7tb0o6DMD2w0XpOEma4g/Az4h9lmN37r2vCYOS3kHIJmxA9LzMJZa471rTju0LgQtrnrND5pXvL2/LYAmu30Ass74Y+BLw+obMHdnQeTOAaoF3Ei/QDYGX2r6/HN+cKA8ly8cSoq/BNKOftUjSWybobbixAXttc1/RyOnpkM2lGc2fJAHA9i+AXyh2MK4MbGj7pobNziMGTn5ie5cy+fzRhm0ONb3SVm219mXwFiIDfkDTjeRNTfZBBlCNY/svwMclzSsXlN7xKyTlGpflQNI8IhA9iyjhfVXSiZW1Pd4DfEvS2xint6Gina44lJi+20TS5cBTiAmlJGmavyFuFh8HPF3S1oS4ZRMN1w/YfkASklYtk8/VeiWnI6XxfsIsexOyILb3lrQRUVX4kaTVgJVrKsdP8ntVkzvJJvKWmKCRMveiLQeSFhINzz3Bv8cDVzZRmx/T27BomHsbeiiWhM4lJkDnEBeQm3KyMGkDSVcTJbSLe9e7mtpMY2x9G9if6FncldiHt4rtl9W2Nd0oQ07/RWwtECFC+le2q8udSHonsULoSbY3kbQZ8HlX2E/XJhlANYykvYF9iCmDy/qeWhN4ZNheMF0g6Tpge9sPlMezgZ/mBOPyk8F60hWSfmJ7bv9rsKeI37DdnYG1iRH5h5q0NR1Q35LpyY5VsnUtsANwVdNBdZNkCa95riCi+nWA4/qO3wss7MSj4eNk4KpydwkxrZHTY1PjAkmvITRs8q4paZMbJO1DLDHeDDiYuC5Wo9xUHUgM6lwHnNRk78s05T7FyrFvEKWvvRnR0KrNg2USHXhUh2/orkuZgWqRUvPdzHYjNd/pjKRtgReWh5c1MIU3rekTGHwYeIBce5K0RNnGcDjwUuJ190PgmF5GuZKNM4DFRJZ/T+AW2/Mm/1tJP5I2Bv6FuM4auBw4xPZvG7D1SeAuopn8IGK91CLbh0/6FweMDKBaYrrUfLui7Jpal9G7jH7XnUdJkgwKYwSLVya2PAydUOdMofRlvp3RQfWXhi07niW89ngPpeYLUFa7/K9uXRoOJB0EHAH8iZAzEHGH1GgPxXRC0gVjg/XxjiVJLSSdPdnzlafwHh2IKBpnFU89M5D0TOAEYF3bW0jaEnil7X+sbcv2I2Utznds31b7/G2RAVR7TIuab0fMA+bYvqNrR4aN0huyOrBOUWjufbKsBazfmWPJTGBH4FbgdOLGscmoZitJ95SfBaxWHmepevn5IvB+4AsAthcWDa9qAZTiA/AI4L3EnkKKoO/xto+uZactMoBqj0skfZB4Y+9B1Hy/17FPw8KtpOjjY+UAYqT7qSytBv25TjxKZgrrAXsQzcj7AOcCp9u+obYh27Nqn3MGsrrtBWOyd7VX7ryP6LHa3vbNAJKeAZwg6X22P1PZXqNkD1RLTJeab5tIOrT8+BxCv+hcRu8y+nQXfg0TRaz198BrbR8vaT/gNcBvgSNt/3eX/iUzg7IAfG9iF+hRtjN4HzDKuqz3Amfa3kbSa4G3296zoo1rgD1s3z7m+FOA84ZNaiUDqGRgkXTEJE97GFO+bSPp58Dutv9b0ouJEeWDgK2BZ9tONfKkMUrgtBcRPG1MqOF/uSzlTQaIkgk6EXgBIUB6M7BvzRUvkq63vcVUnxtUsoTXEpJeSCw13Ij4d+/V5p/RpV+DjO2jACS9zvaZ/c9Jel03Xg0ds/qyTG8ATrR9FnBWEbNLkkaQdCqwBfB9Iut0fccuJRNQKiTb2d69bHpYqSGJnckETYdO7DQzUC0h6Uai/ns1MUkGQDZGL5sJ1uAsdSxZGknXA1uXyaQbgXfZvrT33LDd8SXDg6RHGBFi7P+gycbuAUTSz2xv17CNJYwvzilgtu1VmrRfm8xAtcfdtn/QtRPDhKQ9gZcB60v6bN9Ta1G/uXG6cjoxwHA78BfKOiFJm5KN+UmD2F6pax+SKfEjSf8HOIO+IKdmn+R0a/bPDFRLSPo4MAv4FqMboX/emVMDjqStiF6do4H+hZb3AhfZvrMTx4YMSXOBvyKaNHsLmZ8JrJGvv6RpJJ1m+83LOpZ0i6SbxzmcbSaTkAFUS0i6aJzDtr1r684MGZJWIbKlG9q+qWt/kiRZfsaW24sG3kLbm3foVpKsMFnCa4HSoHeC7W927cuQ8jfAp4DHAU+XtDVwdGUl4yRJKiLpMKCnfdcvcvkQMe2VDACS3kQkU04bc/zNwBLbX+/Gs8EnM1At0UaD3nRF0tXArsDFPZ2Q/t1XSZIMLpI+Zvuwrv1IxkfSVcButv9nzPHHA5fa3rYbzwafzEC1R+MNetOYxbbvHqOQm5F/kgwBtg+TtD4jEi6945d251XSxypjgycA2/eV9olkAjKAao83lO/v6TtmIBv0ls0NkvYBZknaDDgYuKJjn5IkWQ7KAM0bgUWMSLgYyABqMFhN0uN7AyY9JK1JtE0kE5AlvGTgkbQ6cDij1+AcY/uBTh1LkmSZSLoJ2NL2g8v8w0nrlMrIbsCBPdVxSRsD/0q0TRzbnXeDTQZQLVGCgEOJSbJ3lUzKHNvndOxakiRJY5Qda68br0yUDAaSDgQOA9Yoh/4H+LjtE7rzavDJAKolJJ1BqJC/xfYWJaC6wvbWHbs2sEg6e7LncwovSQYfSWcBWwEXMFoD7+DOnErGpZTtGG+Ni6T9bJ/SvleDS/ZAtccmtt8gaW8A2/drTFd0shQ7ArcSatpXEeW7JEmGi7PLVzLgLGP/3TwgA6g+MoBqj4ckrUaZHpO0CX13Y8m4rAfsQWxy3wc4Fzjd9g2depUkyXJj+5Ry7Ush3OEmb2DHkLuK2uNIYD7wNElfI9LZH+jUowHH9hLb823vB8wFfg1cLOm9HbuWJMlyIukVwLXE9Q9JWy+rPJ8MJNnvM4bsgWoRSU8mAgEBP7F9e8cuDTySVgX2IrJQGxOlgC/b/s8u/UqSZPmYQAj3ettbdOtZMhUkXdP7/0uCLOG1hKQLbO9GlKHGHkvGQdKpwBbA94GjbF/fsUtJkkyd8YRwH+nKmeQxc3nXDgwaGUA1jKTZwOrAOpKeyEgdeS1g/c4cGw7eRKi2zwMO7rsAi1jEvFZXjiVJstykEO4AI+nQyZ63/enyPVsnxpABVPMcABwCPJWQMehFAfcAn+vKqWHAdvboJcnwcxAhhPsgMVH7Q+CYTj1K+lmzfJ8DbM/IxOQrgAWdeDQkZA9US0g6yPbxXfsxjEg6zfabl3UsSZIkeWxIuhTYqydlUDShzrX94m49G1wyA9USto+X9AKiEbp/oeapnTk1PDyn/4GklYHcEJ4kQ4Ck7YAPsvS1b8uufErGZV3gob7HD5VjyQRkANUSkk4DNiHGefsXamYANQGSDiMuvKtJuqd3mHhjn9iZY0mSTIWvAe8HriObxweZU4EFkr5NXGdfBXylU48GnCzhtYSkXwKbO//Bp4ykj9k+rGs/kiSZOpJ+bHunrv1Ilo2kbYAXETf3l9m+pmOXBprMQLXH9YSy9n917ciwYfswSesDGzG6BHBpd14lSbKcHCHpSyy9C+9b3bmUTMASIktoMlu4TDKAao91gEWSFjD6IpILcZeBpI8DbwQWMbr8mQFUkgw++wPPAlZh5EPZQAZQA4SkecA7gbOIEt5XJZ2Yw08TkyW8lpC083jHbV/Sti/DhqSbgC1t5+7AJBkyJN1ke07XfiSTI2khsKPt+8rjxwNXZrP/xGQGqiUyUFohfkPcvWYAlSTDxxWSNre9qGtHkkkRIxl+ys+5QHgSMoBqGEn3Mv4SxlTTXn7uB66VNLaH4uDuXEqSZDmZS7x/bybev71rX2Y2BouTgavKFB7Aq4GTOvRn4MkSXjLwSNpvvOO2T2nblyRJpoakjcY7bvuWtn1JJkfStsALy8OcwlsGGUAlQ4Gk1YANbd/UtS9Jkiw/ko4DTsoS3uAjaRYhntk/7fy77jwabHLXWDLwSHoFIUA6vzzeWtLZk/+tJEkGhF8CX5R0laQDJa3dtUPJ0kg6CPgTcD5wDnBu+Z5MQGagkoFH0tXArsDFtp9Xjl1ve4tuPUuSZHmRNIeQNNgbuBz4ou2LuvUq6SHp18Dzbd/RtS/DQmagkmFgse27xxxLkbckGRJKaehZ5et24BfAoZK+0aljST+3AmOvs8kk5BReMgzcIGkfYJakzYCDgSs69ilJkuVA0meAlwMXAh+1vaA89Ymi8ZZ0iKRDy4+/AS6WdC6jp50/3YljQ0AGUMkwcBBwOPGmPh34IXBMpx4lSbK8LAQ+1BNoHMMObTuTLMWa5fvvytfjyheML8GTFLIHKkmSJGkUSU8ENgNm947lLsvBQtLrbJ+5rGPJCBlAJQOPpO2ADwIbM3q8NoX4kmTAkfQOYB6wATFNO5dYEbJrp44lo5D0c9vbLOtYMkKW8JJh4GvA+4HryObxJBk25gHbAz+xvYukZwEf7dinpCBpT+BlwPqSPtv31FrAw914NRxkAJUMA7fZTt2nJBlOHrD9gCQkrWr7xiJpkAwGfwB+BrwSuLrv+L3A+zrxaEjIEl4y8EjajdCOGbsL71udOZUkyXJRdqvtDxxC6LndCaxi+2WdOpaMQtIqRFIlNz4sJxlAJQOPpK8S+jE3MFLCs+23dedVkiRTRdLOwNrAfNsPde1PMkLZ+PAp4HG2ny5pa+Bo26/s2LWBJQOoZOCRdJPtTPknyRAhaTZwILAp0b94ku3sqRlQJtj4cJ3t53br2eCSSuTJMHCFpM27diJJkilxCrAdETztCRzXrTvJMhhv40NmWCYhm8iTYWAucK2km4keKBElvJQxSJLBZfNe9kLSScCCZfz5pFty48MUyQAqGQb+pmsHkiSZMot7P9h+WFKXviTLJjc+TJHsgUoGHknHEf0Ti7r2JUmS5UPSEqC3vkXAasD9jGSQ1+rKtySpQWagkmHgl8AXJa0MnAycPk6tPkmSAcL2rK59SJaNpEk19nIKb2IyA5UMDUV8b39CE+py4Iu2L+rWqyRJkuFF0m3ArUTZ7ioiQ/goti/pwq9hIAOoZCiQNAt4ORFAPQ34JrATcJ/tN3bpW5IkybBSrq17EDemWwLnEln+Gzp1bAjIACoZeCR9hgieLiR6oRb0PZcaUUmSJBWQtCoRSB0LHGX7cx27NNBkD1QyDCwEPmT7vnGe26FtZ5IkSaYTJXDaiwieNgY+C3y7S5+GgcxAJUOBpCcCmwGze8dsX9qdR0mSJMOPpFOBLYDvA9+wfX3HLg0NGUAlA4+kdwDzgA2AawlhzStt79qpY0mSJEOOpEcYkZvoDwhSbmIZ5CqXZBiYB2wP3GJ7F+B5wF3dupQkSTL82F7J9prla62+rzUzeJqcDKCSYeAB2w9A1Opt3whk43iSJEklJJ22PMeSEbKJPBkGfi/pCcB3gPMl3Qnc0rFPSZIk04nn9D8owsXbduTLUJA9UMlQIWlnYG1gvu2HuvYnSZJkmJF0GPBBRlbtQPQ/PQScaPuwrnwbdDKASgYWSbOBA4FNgesIDaiHu/UqSZJk+iHpYxksTY0MoJKBRdIZxEb3y4A9iSbyed16lSRJMj2RtD6wEX3tPSkXMzEZQCUDi6TrbD+3/LwysMD2Nh27lSRJMu2Q9HHgjcAiYEk57FwmPDHZRJ4MMot7P9h+WNJkfzZJkiR57PwtMMf2g107MixkAJUMMltJuqf8LGC18jgF3pIkSeryG2AVIAOo5SQDqGRgsT2rax+SJElmCPcD10q6gL4gyvbB3bk02GQAlSRJkiTJ2eUrWU6yiTxJkiRJEiStBmxo+6aufRkGcpVLkiRJksxwJL2CWNY+vzzeWlJmpCYhA6gkSZIkSY4EdqAsard9LfCMLh0adDKASpIkSZJkse27xxx7pBNPhoRsIk+SJEmS5AZJ+wCzJG0GHAxc0bFPA01moJIkSZIkOQh4DiFhcDpwD3BIpx4NODmFlyRJkiRJMkWyhJckSZIkMxxJ2wEfBDZm9DLhLbvyadDJDFSSJEmSzHAk3QS8H7iOvuZx27d05tSAkxmoJEmSJElus526T1MgM1BJkiRJMsORtBuwNzB2F963OnNqwMkMVJIkSZIk+wPPAlZhpIRnIAOoCcgMVJIkSZLMcCTdZHtO134ME6kDlSRJkiTJFZI279qJYSIzUEmSJEkyw5H0S2AT4GaiB0qAU8ZgYjKASpIkSZIZjqSNxjueMgYTk03kSZIkSZIcDJxke1HXjgwL2QOVJEmSJMkvgS9KukrSgZLW7tqhQSdLeEmSJEmSACBpDiFpsDdwOfBF2xd169VgkhmoJEmSJEmQNIvQgnoWcDvwC+BQSd/o1LEBJTNQSZIkSTLDkfQZ4OXAhUQv1IK+51IjahyyiTxJkiRJkoXAh2zfN85zO7TtzDCQGagkSZIkSZD0RGAzYHbvmO1Lu/NosMkMVJIkSZLMcCS9A5gHbABcC8wFrgR27dKvQSabyJMkSZIkmQdsD9xiexfgecBd3bo02GQAlSRJkiTJA7YfAJC0qu0bgWwcn4Qs4SVJkiRJ8ntJTwC+A5wv6U4g17hMQjaRJ0mSJEnyKJJ2BtYG5tt+qGt/BpUMoJIkSZJkhiJpNnAgsClwHaEB9XC3Xg0HGUAlSZIkyQxF0hnAYuAyYE+iiXxet14NBxlAJUmSJMkMRdJ1tp9bfl4ZWGB7m47dGgpyCi9JkiRJZi6Lez9k6W5qZAYqSZIkSWYokpYAvfUtAlYD7i8/2/ZaXfk26GQAlSRJkiRJMkWyhJckSZIkSTJFMoBKkiRJkiSZIhlAJUmSJEmSTJEMoJIkSZIkSaZIBlBJkiRJkiRTJAOoJEmSJEmSKfL/AXFlY0+Hi42TAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "importances_df.plot(kind='bar', figsize=(10, 5), title='Важность факторов')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a name=\"conclusion\"></a>\n",
    "## Вывод"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* В данной работе проведена работа по предсказанию того, откажется ли клиент от услуг оператора сотовой связи или нет\n",
    "* Сначала данные были прочитаны и по ним были сделаны предварительные выводы \n",
    "* После этого была произведена первичная обработка данных - переименованы столбцы, изменены типы данных \n",
    "* Были добавлены 2 новых столбца - целевой признак `Churn`, который означает отказался ли клиент от услуг и `TotalDays`, который обозначает сколько по времени пользователь пользовался услугами\n",
    "* После этого был проведен исследовательский анализ данных - проанализированы количественные и категоральные признаки, проведена попытка найти связь между ними и целевым признаком\n",
    "* Были сделаны следующие выводы: со временем вероятность отказа от услуг снижается и обычно отказывались от услуг те, кто платил больше\n",
    "* После этого была проведена подготовка данных к обучению моделей: удалены ненужные столбцы, применена техника OHE, датафрейм был разделен на обучающую и тестовую выборки, проведено масштабирование количественных признаков и проведен upsample\n",
    "* Затем на данных были обучены 3 модели: LogisticRegression, RandomForestClassifier и CatBoostClassifier\n",
    "* Лучший результат показала модель CatBoostClassifier со значением метрики ROC-AUC в 0.91, что выше, чем значение, требуемое в задании\n",
    "* Самые важные факторы - сколько клиент пользуется услугами и сколько он платит в месяц "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div style=\"background: #cceeaa; padding: 5px; border: 1px solid green; border-radius: 5px;\">\n",
    "    <font color='green'> <b><u>КОММЕНТАРИЙ РЕВЬЮЕРА</u></b>\n",
    "</font>\n",
    "<font color='green'><br>\n",
    "Алексей, уверенная работа, на мой взгляд)<br>\n",
    "Обработка данных, размышления над факторами, многогранная модельная работа, конечные результаты - многое удалось как надо<br> \n",
    "А что совсем хорошо, так это то, что не стал гнаться за точностью любой ценой, даже ценой наружения бизнес-логики. (это я про даты-признаки, многие их вставляют...)<br>\n",
    "\n",
    "Прежде, чем идти дальше, напомню из прошлых наших общений основные моменты, которые бы мне хотелось зафиксировать, как важные для момделирования (что-то ты и так хорошо используешь):<br>   \n",
    "1) F1, Акураси и AUC - отличные метрики, но нужно держать в голове, что это лишь бобщённые характеристики. А в реальной практике бизнесу часто важно знать детали самой ошибки - из каких ошибок/успехов она складывается. И в этом случае очень кстати будут меры полноты и точности. Рассмотри возможность их использования для своих проектов. Только через них можно выйти на финансовые возможности модели. Ведь перед запуском её в бой, будет сравнение стоимости создания и поддержки модели со стомостью \"оставить как есть\" \n",
    "    \n",
    "2) При анализе факторов влияния (их зависиомть между собой и на целевую) происходит по одному из трёх алгоритмов, исходя из типа шкалы. Могут применять следующие графический и аналитический анализы:<br>\n",
    "- колличественная-колличественная: аналитически методод - корреляции, графический - скатерплот.<br>\n",
    "- колличественная-номинальная: аналитически методод - проверка гипотез (стьюдент, маннауитни), графический - боксплот.<br>\n",
    "- номинальная-номинальная: аналитически методод - Хи-квадрат, графический - круговая диаграмма<br>\n",
    "Это я к тому разделу, который посвящён анализу данных. Там ты сделал интересные предположения и чтобы их доказать, стои проверить аналитически. Вот как раз те методы проверки гипотез, что выше написал.<br>\n",
    "    \n",
    "3) Алгоритм нахождения модели может быть таким:<br>\n",
    "- изучаем проблематику бизнеса и самой темы (в данном случае отток). Пытаемся понять суть явления.<br>\n",
    "- генерируем разные факторы влияния - любые.<br>\n",
    "- дизайнируем (прикидываем модели) какие методы решения есть, чтобы решить проблему.<br>\n",
    "- Смотрим, какие данные и в какой шкале есть. Что нужно ещё, что применить выбранный метод.<br>\n",
    "- строим модели и анализируем те фактры, какие сгенерировали ранее.<br>\n",
    "- Думаем над тем, какие причины (факторы) могут ещё влиять на отток.<br>\n",
    "- если после нескольких итераций точность \"замирает\" - значит достигли потолка)<br>\n",
    " \n",
    "4) Проведение анализа важности факторов с помощью команды feature_importances_ (+графический анализ). Отличная вещь намечается: высока я метрика + график рейтинаг факторов по силе влияния на целевой признак). Важнейшая часть любого модельного финала. Если видно, что часть факторов - слабозначимые для модели. Т.е. не значительно влияют на целевую переменную, то их смело можно убирать из модели. Ну и для бизнеса доносить: так мол и так - на эти факторы надо обращать внимание, а эти не влияют на уход клиента.\n",
    "\n",
    "<font color='red'>\n",
    "Есть пару-тройку вопросов для дополнений:<br>\n",
    "1)При моделировании существует проблематика мультиколлинеарности.(когда количественные факторы имеют сильную корреляцию друг с другом - больше 0.8).Есть ли она у нас?  Если есть - надо  решать по понятиям:) (удалить лишнее)<br><br>\n",
    "2) Будет здорово, если построишь график важности факторов.\n",
    "<font color='green'>\n",
    "Однозначно видно, что ты технически сделал хорошую работу.<br>\n",
    "Просто эти два пункта - важный момент в алгоритме моделирования. И хотелось, чтобы они у тебя остались в финале.<br>\n",
    "\n",
    "Спасибо)<br>\n",
    "БУДУ ЖДАТЬ КОММЕНТАРИЕВ."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-success\">\n",
    "Добавил вычисление и графическое ототбражение важности факторов. Также сформулировал на основе этих данных выводы\n",
    "<br>\n",
    "Удалил поле TotalCharges, потому что оно коррелирует с TotalDays\n",
    "</div>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div style=\"background: #cceeaa; padding: 5px; border: 1px solid green; border-radius: 5px;\">\n",
    "    <font color='green'> <b><u>КОММЕНТАРИЙ РЕВЬЮЕРА 2</u></b>\n",
    "</font>\n",
    "<font color='green'><br>\n",
    "Привет, Алексей! Увидел твои доработки. Хорошо удалось ещё поработать!<br>\n",
    "В общем-то, всё по делу. Мне добавить больше нечего, миссия закончена..(<br>\n",
    "\n",
    "Спасибо за этот труд и за совместную работу на других проектах)<br>\n",
    "Искренне желаю покорить ещё не один аналитический Эверест)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.8"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": true,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}